CALCOFI line transect analysis

1 Introduction

This is a working document for the analysis of marine mammal visual line transect data from the CALCOFI surveys 2004-2023. Michaela Alksne provided the data and is helping troubleshoot. The finalized results are due to be used in two papers:

• Temporal trends, led by Simone Bauman-Pickering. This is in some ways an update of Campbell et al. (2015), which looked at 2004-2013.
• Relationship between survey-level density estimates and microbial + small plankton eDNA, led by Trevor Ruiz.

Note

Notes like this are used throughout the text where there are questions for Michaela and Simone.

2 Effort data

The following code reads in the effort data file and creates two tables: - sample_table, containing one line per transect per cruise - region_table, containing one line per cruise (with Area = 1 otherwise later code complains)

Note, we treat treat a day of effort as the sample unit, as was done by Campbell et al. (2015). (This is because there are only 6 transect lines so not really sufficient for reliable variance estimation using a bootstrap.) (Note also that, if we do go back to using the transect as the sampling unit, the Transect column in the effort spreadsheet is closer to (but not exactly equal to) a segment than transect, so it will take some digging to get out the transect IDs.

#Read in the effort data
effort <- read_excel("./fromMichaela/CalCOFI_2021_2023_Effort_OnTransectOnEffortONLY.xlsx", 
                     guess_max = 7000, na = c("", "NA"))

#Prepare tables for Distance analysis
sample_table <- effort %>% 
  filter(SegmentLength_km_ > 0) %>% #Remove records with 0 effort
  mutate(date = date(StartDateTimeLocal)) %>%
  rename(year = Year, season = Season, Region.Label = Cruise,
         segment_length = SegmentLength_km_) %>%
  group_by(Sample.Label = as.character(date)) %>%
  summarise(Effort = sum(segment_length), year = first(year), season = first(season), date = first(date),
            Region.Label = first(Region.Label)) 
region_table <- sample_table %>%
  group_by(Region.Label) %>%
  summarise(Region.Label = first(Region.Label), year = first(year), season = first(season), start_date = first(date), k = n(), L = sum(Effort)) %>%
  mutate(Area = 1)

Here is a copy of the region_table, i.e., with one line per cruise. \(k\) is the number of samples (transects) and \(L\) is the total line length.

region_table %>% kbl() %>% kable_classic(full_width = FALSE)

Region.Label	year	season	start_date	k	L	Area
CC2004-07	2004	summer	2004-07-13	14	1446.8834	1
CC2004-11	2004	fall	2004-11-02	14	931.6503	1
CC2005-01	2005	winter	2005-01-05	12	903.8109	1
CC2005-04	2005	spring	2005-04-16	14	1278.9707	1
CC2005-07	2005	summer	2005-07-02	14	1304.0305	1
CC2005-11	2005	fall	2005-11-05	13	927.4696	1
CC2006-02	2006	winter	2006-02-04	15	1044.4506	1
CC2006-04	2006	spring	2006-04-02	15	1383.5765	1
CC2006-07	2006	summer	2006-07-09	15	1009.4270	1
CC2006-10	2006	fall	2006-10-22	13	1028.8196	1
CC2007-01	2007	winter	2007-01-13	14	949.0365	1
CC2007-04	2007	spring	2007-04-02	11	768.5812	1
CC2007-07	2007	summer	2007-06-29	15	1005.2511	1
CC2007-11	2007	fall	2007-11-03	13	640.6153	1
CC2008-01	2008	winter	2008-01-08	14	790.1258	1
CC2008-03	2008	spring	2008-03-25	10	499.1738	1
CC2008-08	2008	summer	2008-08-14	14	1061.9797	1
CC2008-10	2008	fall	2008-10-15	14	907.0628	1
CC2009-01	2009	winter	2009-01-08	14	980.7786	1
CC2009-03	2009	spring	2009-03-08	13	785.0440	1
CC2009-07	2009	summer	2009-07-15	16	1287.6075	1
CC2009-11	2009	fall	2009-11-07	12	1037.7810	1
CC2010-01	2010	winter	2010-01-13	15	875.1260	1
CC2010-08	2010	summer	2010-07-31	13	1086.1751	1
CC2010-11	2010	fall	2010-10-29	13	774.4909	1
CC2011-01	2011	winter	2011-01-13	16	1052.1534	1
CC2011-04	2004	spring	2011-04-10	15	1238.2236	1
CC2011-08	2011	summer	2011-07-27	14	1341.8873	1
CC2011-10	2011	fall	2011-10-17	13	1044.0281	1
CC2012-02	2012	winter	2012-01-28	7	453.5515	1
CC2012-03	2012	spring	2012-03-25	8	782.0651	1
CC2012-07	2012	summer	2012-07-03	16	1480.9625	1
CC2012-10	2012	fall	2012-10-20	14	950.4777	1
CC2013-01	2013	winter	2013-01-11	14	848.6145	1
CC2013-04	2013	spring	2013-04-07	10	590.5558	1
CC2013-07	2013	summer	2013-07-07	14	1326.3615	1
CC2013-11	2013	fall	2013-11-10	14	926.1534	1
CC2014-02	2014	winter	2014-01-29	6	457.5277	1
CC2014-04	2014	spring	2014-04-03	11	827.0975	1
CC2014-07	2014	summer	2014-07-06	16	1256.0396	1
CC2014-11	2014	fall	2014-11-09	14	1156.5048	1
CC2015-01	2015	winter	2015-01-16	14	1121.4828	1
CC2015-04	2015	spring	2015-04-05	13	1133.8400	1
CC2015-07	2015	summer	2015-07-08	14	1297.2557	1
CC2015-11	2015	fall	2015-10-29	14	1067.4050	1
CC2016-01	2016	winter	2016-01-08	14	875.7938	1
CC2016-04	2016	spring	2016-04-02	12	1026.1933	1
CC2016-07	2016	summer	2016-07-11	12	1154.7612	1
CC2016-11	2016	fall	2016-11-07	12	849.3831	1
CC2017-01	2017	winter	2017-01-06	13	1027.7116	1
CC2017-04	2017	spring	2017-03-29	13	1265.2516	1
CC2017-08	2017	summer	2017-08-01	13	1284.8930	1
CC2017-11	2017	fall	2017-11-10	12	946.4949	1
CC2018-02	2018	winter	2018-02-02	8	631.6046	1
CC2018-04	2018	spring	2018-04-06	13	1109.4123	1
CC2018-06	2018	summer	2018-06-09	14	1198.7060	1
CC2018-10	2018	fall	2018-10-15	13	1082.9915	1
CC2019-02	2019	winter	2019-02-07	5	370.8165	1
CC2019-04	2019	spring	2019-04-02	11	920.8476	1
CC2019-07	2019	summer	2019-07-12	13	1251.9319	1
CC2019-11	2019	fall	2019-11-05	11	1072.9128	1
CC2020-01	2020	winter	2020-01-05	14	960.1047	1
CC2021-07	2021	summer	2021-07-19	12	1258.6472	1
CC2021-11	2021	fall	2021-11-01	11	826.4556	1
CC2204SH	2022	spring	2022-04-12	8	576.5156	1
CC2208	2022	summer	2022-08-14	13	790.1335	1
CC2211SR	2022	fall	2022-11-05	11	705.2526	1
CC2301RL	2023	winter	2023-01-07	9	594.6193	1
CC2303SH	2023	spring	2023-03-25	13	801.7941	1
CC2311SR	2023	fall	2023-11-05	14	929.0986	1

One thing that’s apparent is that the amount of survey effort, both in terms of line length and days, varies greatly between years. This is highlighted in the following plot, where the red line shows line length and the blue line days of surveying. The black dots are on the winter surveys, which tend to be sorter but not always.

ggplot(region_table, aes(x = start_date)) +
  geom_line( aes(y = L), color = "red") + 
  geom_line( aes(y = k * 100), color = "blue", linetype = "dashed") +
  geom_point( aes(y = L), size = ifelse(region_table$season == "winter", 2, 0)) +
  scale_y_continuous(
    name = "Line length",
    sec.axis = sec_axis(~ . / 100, name="Days of surveying"),
    limits = c(0, 1600)
  ) + 
  xlab("Start date of survey") +
  theme(
    axis.title.y = element_text(color = "red"),
    axis.title.y.right = element_text(color = "blue")
  ) +
  ggtitle("Line transect effort per survey")

If surveys do not cover the whole study area, inference about regional density may be biased if density varies over the region.

Note

Should we look into this further? It may be that we have to remove some surveys?

Note also that there are not necessarily 4 surveys in each year, so if we are taking yearly averages to make inferences about trends we will need to take this into account. (An alternative would be a temporal trend model, with a main effect for season (or time-season interaction).)

One other minor thing: the protocol for naming cruises changed in spring 2022. Before that they had a 4 digit year, then a dash then the month the survey started; after that they have a 2-digit year and then 2 letters afterwards except for summer. No big deal I think, just thought it’s worth noting.

3 Sightings data

In this section we read in the sightings data; we create and check the columns we need for distance sampling detection function analysis.

Note

This is the corrected sightings data that Michaela sent on December 10th 2024.

#Read in sightings data
#Increase guess_max to more than number of rows so it gets the column types right
sightings <- read_excel("./fromMichaela/CalCOFI_2004_2023_CombinedSightings_add_weather_corrected.xlsx", 
                        guess_max = 7000, na = c("", "NA"))

The following variables will be needed for the distance sampling analysis:

• perpendicular distance – see Section 3.1
• whether on effort and on transect - see Section 3.2
• transect id (i.e., (now that we are using date as the sample unit) the date the sighting was made on) - see Section 3.3
• species - will use the Species1 column - see Section 3.4
• school size - will use the best column - see Section 3.5
• other potential covariates for the detection function - see Section 3.6

3.1 Perpendicular distances

Perpendicular distances are required for all observations. These were not supplied for 2022 and 2023 cruises, so will be calculated below. Before this, we check for the records before 2022 that there are perpendular distances for each record:

#Create an indicator for whether the record is 2022 or later
sightings$Year <- year(sightings$Date_Time_Local_)
ind <- sightings$Year >= 2022

#Count the number of records from before 2022 that have missing perpendicular distances
ind2 <- !ind & is.na(sightings$perp_dist_m)
sum(ind2)

[1] 0

There are 0 records with missing perpendicular distances prior to 2022. So no problem there.

Now to obtaining perpendicular distances for 2022 and 2023 cruises. We use the binocular reticle readings (field BinoReticle), observer height and earth radius to calculate radial along-water distance (using formulae in Lerczak and Hobbs 1997). In some cases, the field distance_m was filled, which we take to mean that a radial distance was estimated directly (by eye) so we use this in place of any reticle measurement (in the cases we checked the reticle field was 0).

We then use ships bearing (ShipHeadingTru) and observaton bearing (SightingBearingTru) to get the observation angle and hence perpendicular distance.

#Warnings suppressed - they arise in cases where there is no reticle observation or 
# sighting bearing (which can be the case when a direct radial distance is recorded)

#Get BinoReticle to radians conversion
sightings$BinoReticle <- as.numeric(sightings$BinoReticle)
sightings$ret_radians <- as.numeric(sightings$ret_radians)
ret_to_rad <- sightings$ret_radians[1] / sightings$BinoReticle[1]
sightings$ret_radians2 <- sightings$BinoReticle * ret_to_rad

radius_earth <- sightings$radius_earth[1]

#The following formulae are from Lerczak and Hobbs (1997)
#Get angle from horizontal to horizon
sightings$alpha <- atan(sqrt(2 * sightings$ObsHeight_m_ * radius_earth + sightings$ObsHeight_m_ ^ 2) / radius_earth)
#Get angle from marine mammal to platform
sightings$beta <- pi / 2 - sightings$alpha - sightings$ret_radians2
#Get distance along earth's surface to horizon
sightings$H <- sightings$alpha * radius_earth
#Get line-of-sight distance to marine mammal
sightings$D0 <- (radius_earth + sightings$ObsHeight_m_) * cos(sightings$beta) - 
  sqrt((radius_earth + sightings$ObsHeight_m_) ^ 2 * (cos(sightings$beta)) ^ 2 - 
       (2 * sightings$ObsHeight_m_ * radius_earth + sightings$ObsHeight_m_ ^ 2))
#Get central arc from marine mammal to platform
sightings$delta <- asin(sin(sightings$beta) * sightings$D0 / radius_earth)
#Get distance to marine mammal along the surface of the ocean
sightings$D <- sightings$delta * radius_earth

#If Distance_m_ column is filled in, use that instead
sightings$D2 <- sightings$D
ind2 <- sightings$Distance_m_ > 0
ind2[is.na(ind2)] <- FALSE
sightings$D2[ind2] <- sightings$Distance_m_[ind2]

#Get horizontal angle (in degrees) from trackline
tmp <- sightings$ShipHeadingTru - as.numeric(sightings$SightingBearingTru)
#Trim clearly wrong values
tmp[abs(tmp) > 360] <- NA
#Transform into the 0-90 degree quadrat (will be useful later for looking at forward distances)
sightings$angle2 <- asin(abs(sin(tmp/360 * (2 * pi)))) * 360 / (2 * pi)

#Convert to perpendicular distance
sightings$perp_dist_m2 <- sightings$D2 * abs(sin(sightings$angle2 / 360 * 2 * pi))

One check that can be performed on the above is to check the calculated radial distances using Lerczak and Hobbs against those given in the sightings spreadsheet for sightings earlier than 2022. Here is a histogram of the distance differences:

ind3 <- sightings$dist_calc_m_ > 0
hist(sightings$D[ind3] - sightings$dist_calc_m_[ind3], xlab = "Difference in calculated radial distances (m)", main = "Difference between spreadsheet and Lerczak & Hobbs", breaks = 100)

As can be seen, in the great majority of cases the difference is very small - and even the larger distances are not great, so no need to investigate further.

One other thing to note is that there are 441 records that don’t have a valid number in either the ShipHeadingTru or the SightingBearingTru column of the sightings spreadsheet.

Note

Worth investigating further?

We now check whether any 2022/2023 data still have missing distances:

#Calculate how many missing distances we have from 2022 onwards
ind3 <- ind & is.na(sightings$perp_dist_m2)
tmp <- sightings[ind3, ]       

There are 2 of these. Here is some info about them:

sightings[ind3, c("Cruise", "Date_Time_Local_", "UniqueID")]

# A tibble: 2 × 3
  Cruise   Date_Time_Local_    UniqueID   
  <chr>    <dttm>              <chr>      
1 CC2211SR 2022-11-05 17:11:09 CC2211SR_1 
2 CC2303SH 2023-04-19 11:36:00 CC2303SH_86

Note

Should we investigate further?

For now I will drop these records …

sightings <- sightings[!ind3, ]

… and copy over the newly calculated perp distances for 2022 onwards into the perp_dist_m field used by all the other records.

#Copy the perp_dist_m2 for these records to perp_dist_m
ind <- sightings$Year >= 2022
sightings$perp_dist_m[ind] <- sightings$perp_dist_m2[ind]

3.2 Getting on-effort on-transect data

We wish to select only sightings data from when observers were on effort and the vessel was on transect. There are various potentially relevant fields here; I checked with Michaela and she recommended to use segID_NaN_OffEffortAnd_orOffTransect__AG_, which contains a segment ID for each sighting that was recorded on effort and on transect.

#Select data where OnEffort and OnTransect are ON
ind <- !is.na(sightings$segID_NaN_OffEffortAnd_orOffTransect__AG_)

Of the 6279 sighting records, the number made when both on effort and on transect according to the above criteria are 2951. We select just these data for futher processing.

sightings <- sightings[ind, ]

Note

This is very close to the numbers Michaela gave me in her email of 10th December.

This is a high level of loss of observations with perpendicular distances. We could potentially use off transect observations for fitting the detection function if the crew were on effort – but this risks compromizing pooling robustness (among other things). It will be useful to have a conversation about this.

3.3 Getting transect ID for each sighting

In the original version of this code, I used transect number as the sample unit. Transect number werenot provided with the sightings, so I had written some code to obtain them based on the time of the sighting (the sightings spreadsheet has a column Date_Time_Local_ and the effort spreadsheet has columns StartDateTimeLocal and EndDateTimeLocal). However, Michaela pointed out the sightings table column segID_NaN_OffEffortAnd_orOffTransect__AG_, which gives the segment ID and can be matched with SegmentID in the effort table. So I changed the code to use this to pull in the transect name from the effort table (and the Cruise ID renamed Region.Label; we have to copy it over because although Cruise is recorded in the sightings spreadsheet it is formatted differently).

I have now changed to use date as the sample unit, so could just use the date of the sighting. However, I keep the current merge code, because

• we still need Cruise ID
• it may be that a sighting occurs on a different date to the start of a segment, if the segment goes on for a long time (although this is very unlikely since visual surveys do not take place overnight)

tmp <- effort %>% select(SegmentID, Cruise, StartDateTimeLocal) %>%
  rename(Region.Label = Cruise) %>%
  mutate(date = date(StartDateTimeLocal))
sightings <- left_join(sightings, tmp, by = join_by(segID_NaN_OffEffortAnd_orOffTransect__AG_ == SegmentID), keep = TRUE)

Check for any records that did not merge successfully

ind <- is.na(sightings$StartDateTimeLocal)

There is 1. Here is some information about it:

sightings[ind, c("Cruise", "Date_Time_Local_", "UniqueID")]

# A tibble: 1 × 3
  Cruise Date_Time_Local_    UniqueID  
  <chr>  <dttm>              <chr>     
1 CC0602 2006-02-24 12:06:00 CC0602_076

Note

Need to check with Michaela

I will delete it for now.

sightings <- sightings[!ind, ]

We create a field Sample.Label which (with Region.Label) will be used to join the effort data to the sightings data in the Distance analysis later. Sample.Label is the cruise code, a dash and then the transect number.

sightings <- sightings %>% mutate(Sample.Label = as.character(date))

3.4 Species

Here is a printout of the sample size (of groups) in the sightings data, by species. The table doesn’t look fancy as the standard table formatting (using kbl) didn’t work with some of the characters in the cetacean codes CSV file (there are some odd line feed characters). Note also some codes are missing - but the general idea is there.

#Get species names
sp_names <- read.csv("./fromMichaela/CalCOFI_cetacean_codes.csv")

#Tabulate sightings and add in species names
sightings$Species1 <- toupper(sightings$Species1)
tmp <- tibble(data.frame(sort(table(sightings$Species1), decreasing = TRUE))) %>%
  rename(Code = Var1) %>% 
  left_join(sp_names, by = join_by(Code == CalCOFI.NOAA.Species.Abbreviation))
print(tmp, n = Inf)

# A tibble: 31 × 3
   Code    Freq Common.Name                 
   <chr>  <int> <chr>                       
 1 DD       562 Short-beaked common dolphin 
 2 ULW      556 Unidentified Large Whale    
 3 DSP      316 Unidentified common dolphin 
 4 BP       260 Fin Whale                   
 5 MN       215 Humpback Whale              
 6 UD       175 Unidentified Dolphin        
 7 BM       162 Blue Whale                  
 8 DC        92 Long-beaked common dolphin  
 9 PD        92 Dall's porpoise             
10 LO        81 Pacific white-sided dolphin 
11 ER        73 Gray Whale                  
12 GG        69 Risso's dolphin             
13 TT        65 Bottlenose dolphin          
14 UC        60 Unidentified cetacean       
15 PM        50 Sperm Whale                 
16 BA        25 Minke Whale                 
17 LB        23 Northern right-whale dolphin
18 OTH       19 <NA>                        
19 OO        12 Killer Whale                
20 USC       11 Unidentified�small cetacean 
21 SC         9 Striped dolphin             
22 ZICA       9 Cuvier's beaked whale       
23 BBO        2 <NA>                        
24 GM         2 Short-finned pilot whale    
25 UBW        2 <NA>                        
26 USW        2 Unidentified�small whale    
27 BB         1 Sei Whale                   
28 BEBA       1 �Baird's beaked whale�      
29 PC         1 False Killer whale          
30 UNZIPH     1 Unidentified ziphid         
31 UO         1 Unidentified odontocete     

For the purposes of the analyses undertaken here, we may be interested only in blue, fin and humpback whale; for the trends paper we will likely be interested in more species – but will want some sample size cutoff. Note also we will need to determine how to deal with the unidentified whale sightings (there are many, because the survey was undertaken in passing mode, with relatively inexperienced observers).

3.5 School size

In this section, we check a school size (column Best) is present for all sightings.

ind <- (is.na(sightings$Best))

There are 2 records with missing school size. These are removed in the code below.

Note

I checked with Michaela and she could not locate the school size so OK to drop.

sightings <- sightings[!ind, ]

Note that there are some very large school sizes, and we will need to return to examine these before a sensible analysis can be done on the species involved.

summary(sightings$Best)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
   0.00    1.00    4.00   38.87   21.25 4600.00 

Here is a summary of the number of records of species with school sizes more than 100:

which_ind <- which(sightings$Best > 100)
sort(table(sightings[which_ind, ]$Species1), decreasing = TRUE)

 DD DSP  DC  UD  LB  GG  TT  LO OTH  SC 
106  50  37  19   3   2   2   1   1   1 

and more than 1000:

which_ind <- which(sightings$Best > 1000)
sort(table(sightings[which_ind, ]$Species1), decreasing = TRUE)

 DD  DC  UD DSP 
  6   4   2   1 

These are all one of the two species of common dolphin, or UID dolphin.

3.6 Other covariates for the detection function

In Campbell et al. (2015), they used: “Beaufort sea state (numerical: 0–5), ship (categorical), swell (numerical: wave height in m), school size (numerical), and, for short-beaked common dolphins, school size class (categorical).” We have already looked at school size, but we consider the others (and some alternatives) below.

In addition, it was suggested that observer would be a potentially relevant covariate, but there are too many for this to be possible; this is also true for cruise. Another possible covariate not considered by Campbell et al. is season. It could be argued that some of the sightability variation here would be captured by sea state and swell – but there are potentially animal behavioural (and other) factors that could be included with season so it seems worth adding.

Cue is also recorded as a variable, and could be useful - but there are many missing values so we do not pursue it here.

sum(is.na(sightings$Cue))

[1] 645

There are also a suite of observational variables such as Quality, Visiblity, Precipitation, Could, Glare - but these all have many missing values.

3.6.1 Sea state and swell

For sea state and swell, in an initial round of exploratory analysis we found many missing records. On investigation, Michaela found there were not contemporaneous records for these, and filled in the values with the closest (in time) available record.

Note

Note to self - check if these are both Beaufort and swell or just Beaufort that were filled in by Michaela.

Check there are no missing records:

#Look at beaufort and swell
sum(is.na(sightings$Bft))

[1] 0

sum(is.na(sightings$Swell_ft_))

[1] 0

We were advised to consider truncating records too far away in time. Here is a histogram of the time delays (note that only records that were filled in are shown here, as the others have a Time_Diff_minutes of NA).

hist(sightings$Time_Diff_minutes/60, main = "Time delay between sighting and sea conditions", xlab = "Time delay (hours)", breaks = 100)

There are a small number that are very large and will clearly need truncated – here is a histogram with these removed.

ind <- abs(sightings$Time_Diff_minutes) <= 6000
hist(sightings$Time_Diff_minutes[ind]/60, main = "Time delay between sightings and sea conditions (truncated)", xlab = "Time delay (hours)", breaks = 100)

Following Michela’s suggestion, we will truncate any records with a time delay of more than 8 hours.

ind <- abs(sightings$Time_Diff_minutes) > 8 * 60
ind[is.na(ind)] <- FALSE

This will result in truncating 76 out of 2948 remaining records. Note that some of these are sightings with small perpendicular distances, so it is a shame to truncate them as they will affect the analysis. However, it seems the best choice in the circumstances.

sightings <- sightings[!ind, ]

There are the frequency of the different Beaufort values – note this is in the sightings database, so it does not reflect the proportion of time surveying in the different conditions if there are fewer sightings in higher Beauforts (which may reflect that there was little surveying done in such Beauforts, and/or that animals are hard to sight in those conditions):

table(sightings$Bft)

  0   1   2   3   4   5   6 
 13 233 713 840 567 493  13 

and here is a histogram of swell height

hist(sightings$Swell_ft_, xlab = "Swell height (ft)", main = "Swell height in sightings records")

One thing to check about Beaufort and swell is whether they are strongly correlated – if so it may not be wise to try them both as covariates in the detection function modelling. However (and perhaps surprisingly), the correlation is not very strong:

cor(sightings$Bft, sightings$Swell_ft_)

[1] 0.5651643

3.6.2 Ship / observer height

Regarding the possible ship covariate, we do not have this in the dataset – but we do have observer height, which is strongly related and potentially relevant.

Note

Check if we don’t have ship?

First to check that observer height is not missing for any records:

sum(is.na(sightings$ObsHeight_m_))

[1] 0

Looking at the distribution of observer heights, there are only a few values and so it does not seem suitable for a continuous numerical covariate:

hist(sightings$ObsHeight_m_, breaks = 100, xlab = "Observer height (m)", main = "Observer height in sightings records")

It seems reasonable to group them into three classes: 9m or less, 9-12m and more than 12m (coded as A, B and C).

Note

Check with Simone and Michaela if this is reasonable, as they likely know the sighting conditions associated with the different ships at these heights.

sightings$fObsHeight <- as.factor(ifelse(sightings$ObsHeight_m_ <= 9, "A",
                               ifelse(sightings$ObsHeight_m_ <= 12, "B", "C")))

Note that Distance does not allow ordinal classes, so we won’t be able to constrain estimated sightability so that the detecability of C is more than B and this is more than A. We will have to see how things come out in the estimates.

3.6.3 Season

We did consider this as a covariate, but after exploratory modelling dropped it.

4 Exploratory analysis of distances

After the various data selection noted above, there are a total of 2872 records remaining in the sightings data. Here we explore some aspects of the distance data, to try to identify any issues that might affect the detection function modelling.

First, we plot a histogram of the distances (pooled across species)

hist(sightings$perp_dist_m, breaks = 200, xlab = "Perpendicular distance (m)", 
     main = "Perpendicular distances")

We see a large peak at the lowest distance bin, and a very long right tail, out to 12km. It is typical to truncate the data (in Campbell et al. this was at 2400m for large cetaceans and 600m for small ones except Dall’s porpoise where 700m was used); to focus the plots we truncate off the largest 10% of distances.

trunc_90 <- ceiling(quantile(sightings$perp_dist_m, 0.9))
hist(sightings$perp_dist_m, breaks = 200, xlab = "Perpendicular distance (m)", 
     main = "Perpendicular distances (truncated)",
     xlim = c(0, trunc_90))

The large peak at lowest distances could be due to rounding of perpendicular distances to zero. Because these data are from radial distance and angle, this could come from rounding angles to zero. Of the 2872 sightings, 42 have exactly 0 distance, so it seems unlikely there is a strong problem with rounding of angles to zero. Nevertheless, here is a histogram of the angles we have (recall from earlier that some of the angles are missing), with the quadrants folded so that all angles are between 0 and 90 degrees:

hist(sightings$angle2, breaks = seq(-0.5, 90.5, by = 1), xlab = "Sighting angle (degrees)", main = "Sighting angles")

There is clear evidence of rounding to the nearest 5 degrees, and this could account for some of the spike at zero - but likely not much of it as there are only 42 zero angles.

As an aside, it seems strange to me that we could have rounding of angles to the nearest 5 degrees if the bearing to the animal and ship’s bearing were measured separately – presumably they were not measured separately (or perhaps they were both rounded to the nearest 5 degrees).

Note

Discuss how bearings were obtained, and whether an angle board or similar was used.

More insight could be gained by viewing a 2D map of the sightings (projected into the upper right quadrant). For this we need to know forward as well as perpendicular distances, so we can only do it for the observations where we have an observation angle (note that this is not all the animals). Here is a 2D plot for all of these observations

#Look at forward distances
sightings$forward_distance <- sightings$perp_dist_m / tan(sightings$angle2 / 360 * (2 * pi))
plot(sightings$perp_dist_m, sightings$forward_distance, xlab = "Perpendicular distance (m)", 
     ylab = "Forward distance (m)")

We can clearly see bands that correspond with rounding of radial distances (perhaps acutal distances, or reticle marks). However, our primary interest is in the detections close to zero perpendicular distance. Here we plot all sightings within 100m perpendicular distance:

plot(sightings$perp_dist_m, sightings$forward_distance, xlab = "Perpendicular distance (m)", 
     ylab = "Forward distance (m)", xlim = c(0, 100), 
     ylim = c(0, max(sightings$forward_distance[sightings$perp_dist_m < 100], na.rm = TRUE)))
for(i in 1:5){
  abline(a = 0, b = tan((90 - i) / 360 * 2 * pi), lty = 2)
}

We can see diagonal lines of observations running out at 1 degree, 2 degrees, etc (first 5 degrees out highlighted with dashed lines). This is to be expected and not a problem.

Here is the same plot truncated at 50m perpendicular distance.

plot(sightings$perp_dist_m, sightings$forward_distance, xlab = "Perpendicular distance (m)", 
     ylab = "Forward distance (m)", xlim = c(0, 50), 
     ylim = c(0, max(sightings$forward_distance[sightings$perp_dist_m < 50], na.rm = TRUE)))
for(i in 1:5){
  abline(a = 0, b = tan((90 - i) / 360 * 2 * pi), lty = 2)
}

It looks like many of the small perpendicular distance observations were made at angles close to 90 degrees from the transect line – animals that were almost abeam of the ship.

This can be seen even more clearly when truncating at 20m perpendicular distance:

plot(sightings$perp_dist_m, sightings$forward_distance, xlab = "Perpendicular distance (m)", 
     ylab = "Forward distance (m)", xlim = c(0, 20), 
     ylim = c(0, max(sightings$forward_distance[sightings$perp_dist_m < 20], na.rm = TRUE)))
for(i in 1:5){
  abline(a = 0, b = tan((90 - i) / 360 * 2 * pi), lty = 2)
}

We also see that all of the sightings at 2m or less perpendicular distance were made at small forward distances, except one sighting at 200m forward distance. These detections may have been bow-riding animals.

We make one final 2D plot looking only at animals detected within 20m perpendicular and forward distance:

plot(sightings$perp_dist_m, sightings$forward_distance, xlab = "Perpendicular distance (m)", 
     ylab = "Forward distance (m)", xlim = c(0, 20), 
     ylim = c(0, 20))

We now look into which species are involved. Here we look at the species detected within 20m perpendicular distance:

table(sightings$Species1[sightings$perp_dist_m < 20])

 BA  BM  BP  DC  DD DSP  ER  GG  LB  LO  MN  PD  TT  UD ULW 
  4   3  14   7  75  21   2   2   6  13   9  14   3   4   7 

and within 20m perpendicular distance and 20m forward distance (note this includes only those detections with an angle and therefore a forward distance)

table(sightings$Species1[sightings$perp_dist_m < 20 & sightings$forward_distance < 20])

 BA  BP  DC  DD DSP  GG  LB  LO  MN  PD  UD ULW 
  1   2   3  42   7   1   2   5   1   5   3   2 

Short-beaked common dolphin (DD) dominates with UID common dolphin (DSP) a close second.

However, if we look at perpendicular distances out to (say) 200m by 10m for DD & UID …

ind <- sightings$Species1 %in% c("DD", "DSP")
ind2 <- sightings$perp_dist_m <= 200
hist(sightings$perp_dist_m[ind & ind2], breaks = seq(0, 200, by = 10), 
     xlab = "Perpendicular distance (m)", 
     main = "Pdist (truncated) DD and DSP")

and for the other species in the dataset…

hist(sightings$perp_dist_m[!ind & ind2], breaks = seq(0, 200, by = 10), 
     xlab = "Perpendicular distance (m)", 
     main = "Pdist (truncated) other species")

we see that both have the spike at very small perpendicular distances (0-10m in these plots).

We can try separating all dolphin (and porpoise) species from the others:

ind <- sightings$Species1 %in% c("DD", "DSP", "UD", "DC", "PD", "LO", "GG", "TT", "LB", "USC", "SC")
ind2 <- sightings$perp_dist_m <= 200
hist(sightings$perp_dist_m[ind & ind2], breaks = seq(0, 200, by = 10), 
     xlab = "Perpendicular distance (m)", 
     main = "Pdist (truncated) dolphins")

hist(sightings$perp_dist_m[!ind & ind2], breaks = seq(0, 200, by = 10), 
     xlab = "Perpendicular distance (m)", 
     main = "Pdist (truncated) non-dolphins")

So, although bow-riding may play a part, it seems to be something that affects non-dolphin species to some extent also.

Here we break down by species, for species with more than 50 detections – I set the minimum at 50 rather than 60 because this then include the “UC” (unidentified cetacean) group. The first plot is out to 5000m in increments of 200m, the second to 2500m in increments of 100m, while the second is out to 500m in increments of 20m.

tmp <- table(sightings$Species1)
sp <- names(tmp[tmp>=50])

p <- sightings %>% filter(Species1 %in% sp) %>%
  ggplot (aes(perp_dist_m)) + 
  geom_histogram(breaks = seq(0, 5000, by = 200)) + 
  facet_wrap(~Species1, scales = "free")
p

p <- sightings %>% filter(Species1 %in% sp) %>%
  ggplot (aes(perp_dist_m)) + 
  geom_histogram(breaks = seq(0, 2500, by = 100)) + 
  facet_wrap(~Species1, scales = "free")
p

p <- sightings %>% filter(Species1 %in% sp) %>%
  ggplot (aes(perp_dist_m)) + 
  geom_histogram(breaks = seq(0, 500, by = 25)) + 
  facet_wrap(~Species1, scales = "free")
p

In these plots, we see a spike at small distances in DD, DSP, LO, PD and (to some extent) TT – all dolphin (and related small) species. One exception is MN (humpback), which has 11 sightings within 25m, but then there’s another “spike” at around 500m for this species.

There is more exploratory modelling that could take place, but for now I move on to think about detection function modelling.

One topic to consider is truncation. There is some incentive to use the same truncation distance for groups of species that are grouped under the “unidentified” categories, to make apportioning the UID numbers more straightforward, although it’s not a big deal.

Here is the upper 20th, 10th and 5th quantile by species (for those species with more than 50 detections; sample size also given in the last column):

tmp <- sightings %>% filter(Species1 %in% sp) %>%
  rename(species = Species1) %>%
  group_by(species) %>%
  summarize(q80 = quantile(perp_dist_m, 0.8), q90 = quantile(perp_dist_m, 0.9), q95 = quantile(perp_dist_m, 0.95), n = n())
tmp %>% kbl(digits = 0) %>% kable_classic(full_width = FALSE)

species	q80	q90	q95	n
BM	1377	2323	2760	157
BP	1591	2166	2865	253
DC	716	1464	1787	92
DD	496	767	1121	542
DSP	1154	1680	2316	311
ER	1853	2810	2994	70
GG	969	1284	1883	69
LO	508	683	1023	81
MN	1734	2284	3074	215
PD	668	1067	5349	85
TT	779	1234	2321	65
UC	1778	2411	2982	57
UD	2313	2860	3463	172
ULW	2544	3380	4165	542

Looking first at the large whale species, it looks like the value used by Campbell et al. of 2400m should be ok for these species: BM (blue whale), BP (fin whale), ER (grey whale), MN (humpback). For ULW a larger truncation distance would be suggested, but the issue with UID animals is that their “true” distribution is not uniform as you go away from the transect line (because non-identification gets more likely with increasing distance) so I suggest to use the same as the other large whale species.

Looking second at the smaller species, it looks like the value of 600m may be a rather too small, except for LO (Pacific white-sided dolphin). For the moment, I’ll use:

• 600m for LO (Pacific white-sided dolphin) - update trying 900
• 900m for DC (long-beaked common dolphin), DD (short-beaked common dolphin), DSP (UID common dolphin), PD (Dall’s porpoise)
• 1200m for GG (Risso’s dolphin) and TT (bottlenose dolphin)

For UC (unidentified cetacean) it seems okay, and may be wise, to use the same truncation as the large whale species of 2400m. Lastly, for UD (UID dolphin), the quantiles suggest a truncaton of 2400 again - these are likely large schools seen at the bigger distances, so I’ll stick with 1200m to be congruent with the largest distance used for the other dolphins.

Here is code to implement that:

tmp <- tmp %>%
  mutate(w = c(2400, 2400, 900, 900, 900, 2400, 1200, 900, 2400, 900, 1200, 2400, 1200, 2400))

truncation <- right_join(sightings, tmp, join_by(Species1 == species), keep = TRUE) %>%
  group_by(species) %>%
  summarize(w = first(w), n = first(n), n_after = sum(perp_dist_m <= w), perc_trunc = round((n - n_after) / n * 100))
truncation %>% kbl(digits = 0) %>% kable_classic(full_width = FALSE)

species	w	n	n_after	perc_trunc
BM	2400	157	143	9
BP	2400	253	231	9
DC	900	92	77	16
DD	900	542	497	8
DSP	900	311	214	31
ER	2400	70	62	11
GG	1200	69	60	13
LO	900	81	76	6
MN	2400	215	198	8
PD	900	85	71	16
TT	1200	65	58	11
UC	2400	57	51	11
UD	1200	172	93	46
ULW	2400	542	422	22

5 Detection function modelling by species

5.1 Create observation table

In the code below we pull out from the sightings table just the variables that will be used in the distance analysis. As a check we print a summary of the resulting dataset.

obs_table <- sightings %>%
  rename(sighting_no = SightingNo, date_time = Date_Time_Local_, 
         season = Season, species = Species1, distance = perp_dist_m, ss = Bft, 
         swell = Swell_ft_, size = Best, obs_height = ObsHeight_m_, f_obs_height = fObsHeight, 
         season = Season) %>%
  mutate(date = date(date_time), year = year(date), object = 1:n()) %>%
  select(object, Region.Label, Sample.Label, sighting_no, date, date_time, year, season, species, distance, ss, 
         swell, size, obs_height, f_obs_height)
summary(obs_table)

     object       Region.Label       Sample.Label       sighting_no       
 Min.   :   1.0   Length:2872        Length:2872        Length:2872       
 1st Qu.: 718.8   Class :character   Class :character   Class :character  
 Median :1436.5   Mode  :character   Mode  :character   Mode  :character  
 Mean   :1436.5                                                           
 3rd Qu.:2154.2                                                           
 Max.   :2872.0                                                           
      date              date_time                           year     
 Min.   :2004-07-13   Min.   :2004-07-13 05:34:00.00   Min.   :2004  
 1st Qu.:2009-01-14   1st Qu.:2009-01-14 14:19:00.00   1st Qu.:2009  
 Median :2014-01-31   Median :2014-01-31 15:07:00.00   Median :2014  
 Mean   :2013-11-17   Mean   :2013-11-17 17:22:24.99   Mean   :2013  
 3rd Qu.:2018-02-07   3rd Qu.:2018-02-07 09:14:42.00   3rd Qu.:2018  
 Max.   :2023-11-18   Max.   :2023-11-18 06:41:00.00   Max.   :2023  
    season            species             distance             ss      
 Length:2872        Length:2872        Min.   :    0.0   Min.   :0.00  
 Class :character   Class :character   1st Qu.:  188.0   1st Qu.:2.00  
 Mode  :character   Mode  :character   Median :  553.2   Median :3.00  
                                       Mean   :  949.8   Mean   :3.13  
                                       3rd Qu.: 1245.3   3rd Qu.:4.00  
                                       Max.   :11827.0   Max.   :6.00  
     swell             size           obs_height    f_obs_height
 Min.   : 0.000   Min.   :   0.00   Min.   : 7.10   A:1205      
 1st Qu.: 3.000   1st Qu.:   1.00   1st Qu.: 8.10   B: 876      
 Median : 4.000   Median :   4.00   Median :10.36   C: 791      
 Mean   : 3.956   Mean   :  39.13   Mean   :10.40               
 3rd Qu.: 5.000   3rd Qu.:  22.00   3rd Qu.:13.30               
 Max.   :12.000   Max.   :4600.00   Max.   :17.00               

5.2 Fit detection functions

#Define function to fit detection functions to data passed in as data frame/tibble dat. 
# w is the truncation distance.
# fits a set of key function and adjustments with forward selection of covariates, specified in the function body
# returns a data frame with one row for each fitted detection function, and an AIC column
#Note - season not passed in by default
fit_det_func <- function(dat, w,
  keys = c("unif", "hn", "hr"),
  adjustments = c("cos", "cos", "poly"),
  max_adjustments = c(3, 2, 1),
  covariates = c("ss", "swell", "size", "f_obs_height"),
  covariates_and_adjustments = FALSE){
  
  #Note only fits covariates for hn and hr key functions
  
  n_covariates <- length(covariates)
  model <- 0
  for(key_series in 1:length(keys)){
  
    key <- keys[key_series]
    adjustment <- adjustments[key_series]
    for(nadj in 0:max_adjustments[key_series]){
      
      #Fit model with no covariates - assumes no error will be generated here
      res <- ds(dat, w, key = key, adjustment = adjustment, nadj = nadj, quiet = TRUE)
      model <- model + 1
      min_AIC_within_keyfct <- AIC(res)$AIC
      if(model == 1){
        res_table <- data.frame(model = model, key = key, adjustment = adjustment,
                                nadj = nadj, covar = "", AIC = min_AIC_within_keyfct)
      } else {
        res_table <- rbind(res_table, 
                           data.frame(model = model, key = key, adjustment = adjustment,
                                nadj = nadj, covar = "", AIC = min_AIC_within_keyfct))
      }
      
      if((key %in% c("hn", "hr")) & ((nadj == 0) | covariates_and_adjustments)){
        #Forward selection for model covariates
        round <- 1
        covar_selected <- rep(FALSE, n_covariates)
        while(round <= n_covariates){
          res_AIC <- rep(Inf, n_covariates)
          for(cov in 1:n_covariates){
            #Skip this model if the covariate is already selected
            if(covar_selected[cov]) next
            #Build up the covariate string
            if(any(covar_selected)) {
              covar_text <- paste(paste(covariates[covar_selected], collapse = " + "), covariates[cov], sep = " + ")
            } else {
              covar_text <- covariates[cov]
            }
            #Fit the model
            res <- try(ds(dat, w, key = key, adjustment = adjustment, nadj = nadj,
                      formula = eval(parse(text = paste("~", covar_text))), quiet = TRUE))
            if(inherits(res, "try-error")){
              #There was an error during the fitting - do not save the result, and move on to the next model
              next()
            } else {
              #Save the result
              model <- model + 1
              res_AIC[cov] <- AIC(res)$AIC
              res_table <- rbind(res_table, 
                                 data.frame(model = model, key = key, adjustment = adjustment,
                                      nadj = nadj, covar = covar_text, AIC = res_AIC[cov]))
            }
          }
          #Select best model in this round
          best_model <- which.min(res_AIC)
          if(res_AIC[best_model] < min_AIC_within_keyfct){
            min_AIC_within_keyfct <- res_AIC[best_model]  
            covar_selected[best_model] <- TRUE
            round <- round + 1
          } else{
            break
          }
        }
      }
      
      #Determine whether to fit more adjustments
      if(nadj == 0){
        min_AIC <- min_AIC_within_keyfct
      } else {
        if(min_AIC_within_keyfct < min_AIC){
          #Go on to next adjustment term
          min_AIC <- min_AIC_within_keyfct
        } else {
          #No more adjustments
          break
        }
      }
      
    } #end adjustments
  } #end key_series

  res_order <- order(res_table$AIC)
  res_table <- res_table[res_order, ]
  res_table$Delta_AIC <- res_table$AIC - res_table$AIC[1]
  return(res_table)
}

Run the above function to fit detection functions.

# The above suppresses messages as ds generates lots!

convert_units <- 1/1000000

n_species <- nrow(truncation)
fits <- det_func_tables <- vector("list", n_species)
names(fits) <- names(det_func_tables) <- truncation$species

for(i in 1:n_species){
  
  print(paste("*** Fitting species", truncation$species[i], "***"))
  dat <- obs_table[obs_table$species == truncation$species[i], ]
  
  #Fit the detection functions
  det_func_tables[[i]] <- fit_det_func(dat, truncation$w[i])
  
  #Refit with the AIC_best model
  if(det_func_tables[[i]]$covar[1] == "") {
    fits[[i]] <- ds(dat, truncation$w[i], key = det_func_tables[[i]]$key[1], adjustment = det_func_tables[[i]]$adjustment[1], 
                    nadj = det_func_tables[[i]]$nadj[1], region_table = region_table, sample_table = sample_table,
                    convert_units = convert_units, quiet = TRUE)
  } else {
    fits[[i]] <- ds(dat, truncation$w[i], key = det_func_tables[[i]]$key[1], adjustment = det_func_tables[[i]]$adjustment[1], 
                    nadj = det_func_tables[[i]]$nadj[1], region_table = region_table, sample_table = sample_table,
                    formula = eval(parse(text = paste("~", det_func_tables[[i]]$covar[1]))), 
                    convert_units = convert_units, quiet = TRUE)
  }

}

Print out model selection tables

for(i in 1:n_species) {
  tmp <- det_func_tables[[i]] %>% kbl(digits = 2, caption = paste("Species:", truncation$species[i])) %>% kable_classic(full_width = FALSE)
  print(tmp)
}

Species: BM

	model	key	adjustment	nadj	covar	AIC	Delta_AIC
2	2	unif	cos	1		2169.44	0.00
4	4	hn	cos	0		2170.04	0.60
3	3	unif	cos	2		2170.32	0.88
5	5	hn	cos	0	ss	2170.46	1.02
10	10	hr	poly	0		2170.79	1.35
7	7	hn	cos	0	size	2170.86	1.42
13	13	hr	poly	0	size	2171.11	1.67
6	6	hn	cos	0	swell	2171.96	2.52
9	9	hn	cos	1		2172.03	2.59
14	14	hr	poly	0	f_obs_height	2172.36	2.92
12	12	hr	poly	0	swell	2172.77	3.33
15	15	hr	poly	1		2172.79	3.35
8	8	hn	cos	0	f_obs_height	2173.13	3.69
1	1	unif	cos	0		2226.00	56.56
11	11	hr	poly	0	ss	2231.98	62.54

Species: BP

	model	key	adjustment	nadj	covar	AIC	Delta_AIC
14	14	hn	cos	0	ss + size + f_obs_height	3484.10	0.00
15	15	hn	cos	0	ss + size + f_obs_height + swell	3484.69	0.58
11	11	hn	cos	0	ss + size	3485.89	1.78
13	13	hn	cos	0	ss + size + swell	3486.18	2.07
26	26	hr	poly	0	ss + size + f_obs_height	3491.66	7.55
27	27	hr	poly	0	ss + size + f_obs_height + swell	3492.08	7.98
23	23	hr	poly	0	ss + size	3492.27	8.16
24	24	hr	poly	0	ss + f_obs_height	3492.61	8.50
6	6	hn	cos	0	ss	3492.88	8.77
10	10	hn	cos	0	ss + swell	3492.89	8.78
18	18	hr	poly	0	ss	3493.08	8.98
25	25	hr	poly	0	ss + size + swell	3493.31	9.20
22	22	hr	poly	0	ss + swell	3494.32	10.21
12	12	hn	cos	0	ss + f_obs_height	3494.65	10.54
19	19	hr	poly	0	swell	3496.12	12.01
7	7	hn	cos	0	swell	3496.49	12.38
16	16	hn	cos	1		3500.99	16.89
20	20	hr	poly	0	size	3501.12	17.01
8	8	hn	cos	0	size	3501.86	17.76
17	17	hr	poly	0		3502.39	18.28
21	21	hr	poly	0	f_obs_height	3502.57	18.47
3	3	unif	cos	2		3503.53	19.43
4	4	unif	cos	3		3503.55	19.45
28	28	hr	poly	1		3503.57	19.47
5	5	hn	cos	0		3503.67	19.56
2	2	unif	cos	1		3505.52	21.42
9	9	hn	cos	0	f_obs_height	3507.21	23.10
1	1	unif	cos	0		3595.85	111.74

Species: DC

	model	key	adjustment	nadj	covar	AIC	Delta_AIC
7	7	hn	cos	0	size	1022.79	0.00
4	4	hn	cos	0		1024.37	1.59
10	10	hn	cos	0	size + swell	1024.56	1.78
9	9	hn	cos	0	size + ss	1024.59	1.80
2	2	unif	cos	1		1024.94	2.15
12	12	hn	cos	1		1025.54	2.75
16	16	hr	poly	0	size	1025.91	3.13
11	11	hn	cos	0	size + f_obs_height	1026.26	3.48
5	5	hn	cos	0	ss	1026.34	3.55
6	6	hn	cos	0	swell	1026.37	3.59
13	13	hr	poly	0		1026.65	3.87
3	3	unif	cos	2		1026.68	3.90
21	21	hr	poly	1		1027.06	4.28
19	19	hr	poly	0	size + swell	1027.45	4.67
18	18	hr	poly	0	size + ss	1027.89	5.11
8	8	hn	cos	0	f_obs_height	1028.07	5.29
15	15	hr	poly	0	swell	1028.58	5.79
14	14	hr	poly	0	ss	1028.65	5.86
20	20	hr	poly	0	size + f_obs_height	1029.38	6.60
17	17	hr	poly	0	f_obs_height	1030.30	7.51
1	1	unif	cos	0		1047.57	24.78

Species: DD

	model	key	adjustment	nadj	covar	AIC	Delta_AIC
23	23	hr	poly	0	ss + size	6332.82	0.00
25	25	hr	poly	0	ss + size + swell	6333.89	1.07
18	18	hr	poly	0	ss	6334.06	1.24
22	22	hr	poly	0	ss + swell	6335.50	2.69
26	26	hr	poly	0	ss + size + f_obs_height	6336.50	3.69
24	24	hr	poly	0	ss + f_obs_height	6337.97	5.15
27	27	hr	poly	1		6344.49	11.67
20	20	hr	poly	0	size	6353.87	21.06
19	19	hr	poly	0	swell	6356.56	23.75
17	17	hr	poly	0		6357.30	24.48
16	16	hn	cos	2		6359.55	26.73
21	21	hr	poly	0	f_obs_height	6360.05	27.23
4	4	unif	cos	3		6374.37	41.56
15	15	hn	cos	1		6378.67	45.86
11	11	hn	cos	0	ss + size	6401.80	68.98
13	13	hn	cos	0	ss + size + swell	6402.19	69.37
6	6	hn	cos	0	ss	6402.65	69.83
10	10	hn	cos	0	ss + swell	6402.71	69.90
14	14	hn	cos	0	ss + size + f_obs_height	6403.55	70.73
12	12	hn	cos	0	ss + f_obs_height	6404.42	71.60
7	7	hn	cos	0	swell	6408.62	75.80
3	3	unif	cos	2		6411.85	79.04
8	8	hn	cos	0	size	6415.05	82.23
5	5	hn	cos	0		6416.58	83.77
9	9	hn	cos	0	f_obs_height	6418.80	85.99
2	2	unif	cos	1		6443.43	110.61
1	1	unif	cos	0		6761.58	428.76

Species: DSP

	model	key	adjustment	nadj	covar	AIC	Delta_AIC
22	22	hr	poly	0	swell + size	2868.44	0.00
21	21	hr	poly	0	swell + ss	2868.58	0.14
18	18	hr	poly	0	swell	2869.53	1.09
25	25	hr	poly	0	swell + size + f_obs_height	2869.71	1.28
23	23	hr	poly	0	swell + f_obs_height	2870.95	2.51
11	11	hn	cos	0	ss + size	2871.27	2.84
13	13	hn	cos	0	ss + size + swell	2873.20	4.76
14	14	hn	cos	0	ss + size + f_obs_height	2873.94	5.51
17	17	hr	poly	0	ss	2876.43	7.99
6	6	hn	cos	0	ss	2879.54	11.11
10	10	hn	cos	0	ss + swell	2881.45	13.01
19	19	hr	poly	0	size	2882.18	13.75
12	12	hn	cos	0	ss + f_obs_height	2882.40	13.97
20	20	hr	poly	0	f_obs_height	2883.91	15.47
7	7	hn	cos	0	swell	2884.46	16.02
4	4	unif	cos	3		2886.60	18.16
8	8	hn	cos	0	size	2888.21	19.77
16	16	hr	poly	0		2888.45	20.01
3	3	unif	cos	2		2889.70	21.26
26	26	hr	poly	1		2889.80	21.37
2	2	unif	cos	1		2890.58	22.14
5	5	hn	cos	0		2891.30	22.86
9	9	hn	cos	0	f_obs_height	2892.65	24.22
15	15	hn	cos	1		2900.02	31.58
1	1	unif	cos	0		2911.42	42.99
24	24	hr	poly	0	swell + size + ss	2921.40	52.96

Species: ER

	model	key	adjustment	nadj	covar	AIC	Delta_AIC
9	9	hn	cos	1		951.69	0.00
11	11	hr	poly	0		952.06	0.38
10	10	hn	cos	2		952.60	0.91
4	4	hn	cos	0		952.66	0.98
2	2	unif	cos	1		952.93	1.25
13	13	hr	poly	0	swell	953.28	1.59
6	6	hn	cos	0	swell	953.59	1.91
14	14	hr	poly	0	size	953.84	2.16
3	3	unif	cos	2		953.87	2.18
16	16	hr	poly	1		953.93	2.25
12	12	hr	poly	0	ss	954.04	2.35
5	5	hn	cos	0	ss	954.54	2.86
7	7	hn	cos	0	size	954.60	2.91
15	15	hr	poly	0	f_obs_height	954.80	3.12
8	8	hn	cos	0	f_obs_height	955.48	3.80
1	1	unif	cos	0		965.12	13.43

Species: GG

	model	key	adjustment	nadj	covar	AIC	Delta_AIC
18	18	hr	poly	0	size + ss	836.89	0.00
5	5	hn	cos	0	ss	836.89	0.00
10	10	hn	cos	0	ss + size	838.15	1.26
9	9	hn	cos	0	ss + swell	838.88	1.99
11	11	hn	cos	0	ss + f_obs_height	840.27	3.38
6	6	hn	cos	0	swell	840.85	3.96
19	19	hr	poly	0	size + swell	841.14	4.25
16	16	hr	poly	0	size	841.52	4.63
15	15	hr	poly	0	swell	841.62	4.74
7	7	hn	cos	0	size	842.40	5.51
4	4	hn	cos	0		843.41	6.52
20	20	hr	poly	0	size + f_obs_height	843.91	7.02
2	2	unif	cos	1		843.95	7.06
13	13	hr	poly	0		845.11	8.22
12	12	hn	cos	1		845.18	8.29
3	3	unif	cos	2		845.28	8.39
8	8	hn	cos	0	f_obs_height	846.45	9.56
22	22	hr	poly	1		847.10	10.21
17	17	hr	poly	0	f_obs_height	848.53	11.64
1	1	unif	cos	0		850.81	13.92
14	14	hr	poly	0	ss	855.96	19.08
21	21	hr	poly	0	ss + size + swell	860.80	23.92

Species: LO

	model	key	adjustment	nadj	covar	AIC	Delta_AIC
15	15	hr	poly	0	ss	974.17	0.00
20	20	hr	poly	0	ss + size	975.56	1.39
21	21	hr	poly	0	ss + f_obs_height	976.66	2.49
6	6	hn	cos	0	ss	976.92	2.75
10	10	hn	cos	0	ss + swell	978.09	3.92
11	11	hn	cos	0	ss + size	978.66	4.49
16	16	hr	poly	0	swell	980.17	6.00
12	12	hn	cos	0	ss + f_obs_height	980.61	6.44
14	14	hr	poly	0		985.34	11.17
22	22	hr	poly	1		986.79	12.62
17	17	hr	poly	0	size	987.33	13.16
18	18	hr	poly	0	f_obs_height	987.72	13.55
7	7	hn	cos	0	swell	989.24	15.07
13	13	hn	cos	1		989.26	15.09
4	4	unif	cos	3		990.45	16.28
5	5	hn	cos	0		991.53	17.36
3	3	unif	cos	2		991.79	17.62
8	8	hn	cos	0	size	992.99	18.82
9	9	hn	cos	0	f_obs_height	993.14	18.98
2	2	unif	cos	1		993.46	19.29
1	1	unif	cos	0		1033.96	59.79
19	19	hr	poly	0	ss + swell	1041.96	67.79

Species: MN

	model	key	adjustment	nadj	covar	AIC	Delta_AIC
12	12	hn	cos	0	swell + f_obs_height + ss	3044.17	0.00
2	2	unif	cos	1		3044.53	0.36
3	3	unif	cos	2		3045.23	1.06
11	11	hn	cos	0	swell + f_obs_height	3045.59	1.42
14	14	hn	cos	0	ss + swell + f_obs_height + size	3045.95	1.77
6	6	hn	cos	0	swell	3045.95	1.78
16	16	hr	poly	0		3046.13	1.96
9	9	hn	cos	0	swell + ss	3046.14	1.97
19	19	hr	poly	0	size	3046.38	2.21
10	10	hn	cos	0	swell + size	3046.75	2.58
4	4	hn	cos	0		3046.77	2.60
18	18	hr	poly	0	swell	3047.21	3.04
13	13	hn	cos	0	swell + f_obs_height + size	3047.37	3.19
15	15	hn	cos	1		3047.39	3.22
7	7	hn	cos	0	size	3047.82	3.64
17	17	hr	poly	0	ss	3047.94	3.77
21	21	hr	poly	1		3048.13	3.96
20	20	hr	poly	0	f_obs_height	3048.21	4.04
5	5	hn	cos	0	ss	3048.70	4.53
8	8	hn	cos	0	f_obs_height	3049.06	4.88
1	1	unif	cos	0		3082.16	37.98

Species: PD

	model	key	adjustment	nadj	covar	AIC	Delta_AIC
18	18	hr	poly	0	f_obs_height	911.15	0.00
20	20	hr	poly	0	f_obs_height + swell	912.91	1.76
19	19	hr	poly	0	f_obs_height + ss	913.08	1.93
21	21	hr	poly	0	f_obs_height + size	913.15	2.00
15	15	hr	poly	0	ss	915.74	4.58
14	14	hr	poly	0		917.56	6.41
17	17	hr	poly	0	size	917.79	6.63
16	16	hr	poly	0	swell	918.05	6.90
22	22	hr	poly	1		919.31	8.16
6	6	hn	cos	0	ss	925.36	14.20
13	13	hn	cos	1		925.93	14.77
5	5	hn	cos	0		926.07	14.92
3	3	unif	cos	2		926.17	15.02
11	11	hn	cos	0	ss + size	926.44	15.29
10	10	hn	cos	0	ss + swell	927.36	16.20
8	8	hn	cos	0	size	927.48	16.33
7	7	hn	cos	0	swell	927.54	16.39
4	4	unif	cos	3		927.72	16.57
2	2	unif	cos	1		927.81	16.65
12	12	hn	cos	0	ss + f_obs_height	929.33	18.18
9	9	hn	cos	0	f_obs_height	929.72	18.57
1	1	unif	cos	0		965.94	54.79

Species: TT

	model	key	adjustment	nadj	covar	AIC	Delta_AIC
16	16	hr	poly	0	f_obs_height	778.84	0.00
17	17	hr	poly	0	f_obs_height + ss	779.57	0.73
19	19	hr	poly	0	f_obs_height + size	780.13	1.29
18	18	hr	poly	0	f_obs_height + swell	780.25	1.41
10	10	hn	cos	1		780.66	1.82
12	12	hr	poly	0		781.64	2.80
4	4	unif	cos	3		782.09	3.25
11	11	hn	cos	2		782.42	3.58
20	20	hr	poly	1		782.66	3.82
13	13	hr	poly	0	ss	782.74	3.90
15	15	hr	poly	0	size	782.92	4.08
14	14	hr	poly	0	swell	783.45	4.61
5	5	hn	cos	0		786.03	7.19
3	3	unif	cos	2		786.18	7.34
8	8	hn	cos	0	size	786.51	7.68
7	7	hn	cos	0	swell	787.83	8.99
6	6	hn	cos	0	ss	787.98	9.14
9	9	hn	cos	0	f_obs_height	788.75	9.91
2	2	unif	cos	1		789.17	10.33
1	1	unif	cos	0		822.45	43.61

Species: UC

	model	key	adjustment	nadj	covar	AIC	Delta_AIC
15	15	hr	poly	0	swell	776.63	0.00
7	7	hn	cos	0	size	777.27	0.64
19	19	hr	poly	0	swell + size	777.58	0.95
10	10	hn	cos	0	size + swell	777.69	1.06
18	18	hr	poly	0	swell + ss	778.19	1.56
4	4	hn	cos	0		778.34	1.71
12	12	hn	cos	1		778.45	1.82
11	11	hn	cos	0	size + f_obs_height	778.67	2.04
2	2	unif	cos	1		778.91	2.28
13	13	hr	poly	0		778.92	2.29
9	9	hn	cos	0	size + ss	779.11	2.48
6	6	hn	cos	0	swell	779.91	3.27
20	20	hr	poly	0	swell + f_obs_height	780.09	3.46
21	21	hr	poly	1		780.15	3.52
16	16	hr	poly	0	size	780.22	3.59
5	5	hn	cos	0	ss	780.31	3.68
3	3	unif	cos	2		780.55	3.92
14	14	hr	poly	0	ss	780.87	4.24
8	8	hn	cos	0	f_obs_height	782.18	5.55
17	17	hr	poly	0	f_obs_height	782.57	5.94
1	1	unif	cos	0		793.89	17.26

Species: UD

	model	key	adjustment	nadj	covar	AIC	Delta_AIC
18	18	hr	poly	0	ss + f_obs_height	1304.27	0.00
10	10	hn	cos	0	ss + f_obs_height + swell	1304.62	0.34
8	8	hn	cos	0	f_obs_height + ss	1305.28	1.00
19	19	hr	poly	0	ss + f_obs_height + swell	1306.20	1.93
7	7	hn	cos	0	f_obs_height	1306.86	2.59
5	5	hn	cos	0	ss	1307.31	3.03
9	9	hn	cos	0	f_obs_height + swell	1307.35	3.08
14	14	hr	poly	0	ss	1307.39	3.11
16	16	hr	poly	0	f_obs_height	1308.05	3.77
17	17	hr	poly	0	ss + swell	1308.84	4.57
2	2	unif	cos	1		1308.90	4.62
4	4	hn	cos	0		1309.85	5.58
15	15	hr	poly	0	swell	1310.10	5.83
3	3	unif	cos	2		1310.72	6.44
13	13	hr	poly	0		1311.12	6.84
12	12	hn	cos	1		1311.37	7.09
6	6	hn	cos	0	swell	1311.69	7.42
21	21	hr	poly	1		1313.12	8.84
1	1	unif	cos	0		1318.75	14.48
11	11	hn	cos	0	ss + f_obs_height + size	1328.75	24.48
20	20	hr	poly	0	ss + f_obs_height + size	1330.74	26.47

Species: ULW

	model	key	adjustment	nadj	covar	AIC	Delta_AIC
21	21	hr	poly	0	swell + ss	6523.23	0.00
18	18	hr	poly	0	swell	6524.25	1.02
17	17	hr	poly	0	ss	6524.54	1.32
13	13	hn	cos	0	ss + swell + f_obs_height	6524.56	1.34
9	9	hn	cos	0	swell + ss	6524.73	1.50
6	6	hn	cos	0	swell	6525.14	1.91
24	24	hr	poly	0	ss + swell + size	6525.17	1.95
14	14	hn	cos	0	ss + swell + f_obs_height + size	6525.63	2.41
12	12	hn	cos	0	ss + swell + size	6525.71	2.48
11	11	hn	cos	0	swell + f_obs_height	6525.80	2.58
10	10	hn	cos	0	swell + size	6525.93	2.71
25	25	hr	poly	0	ss + swell + f_obs_height	6526.14	2.92
22	22	hr	poly	0	swell + size	6526.18	2.95
5	5	hn	cos	0	ss	6526.18	2.96
23	23	hr	poly	0	swell + f_obs_height	6526.63	3.40
2	2	unif	cos	1		6531.97	8.74
3	3	unif	cos	2		6533.93	10.71
16	16	hr	poly	0		6534.56	11.33
4	4	hn	cos	0		6535.62	12.40
19	19	hr	poly	0	size	6536.50	13.27
26	26	hr	poly	1		6536.65	13.43
20	20	hr	poly	0	f_obs_height	6537.25	14.03
8	8	hn	cos	0	f_obs_height	6537.40	14.17
15	15	hn	cos	1		6537.40	14.17
7	7	hn	cos	0	size	6537.55	14.33
1	1	unif	cos	0		6569.04	45.82

5.3 AIC-best detection functions

for(i in 1:n_species) {
  cat("\n*** Results for species", truncation$species[i], "***\n")
  print(summary(fits[[i]]$ddf))
  plot(fits[[i]])
  print(gof_ds(fits[[i]]))
}

*** Results for species BM ***

Summary for ds object
Number of observations :  143 
Distance range         :  0  -  2400 
AIC                    :  2169.44 
Optimisation           :  mrds (slsqp) 

Detection function:
 Uniform key function with cosine adjustment term of order 1 

Detection function parameters 
Scale coefficient(s): 
NULL

Adjustment term coefficient(s):  
              estimate         se
cos, order 1 0.8187299 0.07231696


Strict monotonicity constraints were enforced.
                       Estimate         SE         CV
Average p             0.5498343  0.0218627 0.03976234
N in covered region 260.0783751 17.8851043 0.06876813

Strict monotonicity constraints were enforced.

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 0.0788666 p-value = 0.698436

*** Results for species BP ***

Summary for ds object
Number of observations :  231 
Distance range         :  0  -  2400 
AIC                    :  3484.105 
Optimisation           :  mrds (nlminb) 

Detection function:
 Half-normal key function 

Detection function parameters 
Scale coefficient(s): 
                estimate         se
(Intercept)    7.4514577 0.25167634
ss            -0.2560141 0.06033633
size           0.2215301 0.10574033
f_obs_heightB -0.1780100 0.15026538
f_obs_heightC  0.3190190 0.20235809

                       Estimate          SE         CV
Average p             0.4913724  0.02738609 0.05573387
N in covered region 470.1118759 35.01815952 0.07448899

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 0.181474 p-value = 0.306431

*** Results for species DC ***

Summary for ds object
Number of observations :  77 
Distance range         :  0  -  900 
AIC                    :  1022.785 
Optimisation           :  mrds (nlminb) 

Detection function:
 Half-normal key function 

Detection function parameters 
Scale coefficient(s): 
               estimate          se
(Intercept) 5.803332000 0.154790655
size        0.002097156 0.001782221

                       Estimate          SE         CV
Average p             0.5540501  0.05141113 0.09279149
N in covered region 138.9766134 16.91552515 0.12171490

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 0.0764608 p-value = 0.712573

*** Results for species DD ***

Summary for ds object
Number of observations :  497 
Distance range         :  0  -  900 
AIC                    :  6332.816 
Optimisation           :  mrds (nlminb) 

Detection function:
 Hazard-rate key function 

Detection function parameters 
Scale coefficient(s): 
                estimate           se
(Intercept)  5.946395288 0.3814389814
ss          -0.513964221 0.1075166709
size         0.001084942 0.0008447454

Shape coefficient(s):  
             estimate         se
(Intercept) 0.1428196 0.08081657

                        Estimate           SE        CV
Average p              0.1904508   0.02155415 0.1131744
N in covered region 2609.5984630 314.63879352 0.1205698

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 0.243021 p-value = 0.197568

*** Results for species DSP ***

Summary for ds object
Number of observations :  214 
Distance range         :  0  -  900 
AIC                    :  2868.436 
Optimisation           :  mrds (nlminb) 

Detection function:
 Hazard-rate key function 

Detection function parameters 
Scale coefficient(s): 
               estimate         se
(Intercept)  7.39394787 0.72027875
swell       -0.61642778 0.16786993
size         0.02301244 0.01130829

Shape coefficient(s):  
            estimate        se
(Intercept)        0 0.1862348

                       Estimate           SE        CV
Average p             0.3356156   0.09605746 0.2862127
N in covered region 637.6341793 187.45575498 0.2939864

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 0.0937552 p-value = 0.616364

*** Results for species ER ***

Summary for ds object
Number of observations :  62 
Distance range         :  0  -  2400 
AIC                    :  951.6852 
Optimisation           :  mrds (slsqp) 

Detection function:
 Half-normal key function with cosine adjustment term of order 2 

Detection function parameters 
Scale coefficient(s): 
            estimate        se
(Intercept) 7.153972 0.1521713

Adjustment term coefficient(s):  
              estimate        se
cos, order 2 0.3089957 0.1742535


Strict monotonicity constraints were enforced.
                       Estimate          SE        CV
Average p             0.4767994  0.07334275 0.1538231
N in covered region 130.0337225 23.29754979 0.1791654

Strict monotonicity constraints were enforced.

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 0.0220906 p-value = 0.994592

*** Results for species GG ***

Summary for ds object
Number of observations :  60 
Distance range         :  0  -  1200 
AIC                    :  836.8871 
Optimisation           :  mrds (nlminb) 

Detection function:
 Hazard-rate key function 

Detection function parameters 
Scale coefficient(s): 
                estimate          se
(Intercept)  7.212901751 0.276343227
size         0.006137863 0.007925481
ss          -0.253455344 0.083559224

Shape coefficient(s):  
            estimate        se
(Intercept) 1.859662 0.6460888

                      Estimate        SE         CV
Average p            0.7052541 0.0556863 0.07895921
N in covered region 85.0757235 9.2432427 0.10864724

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 0.0459515 p-value = 0.900366

*** Results for species LO ***

Summary for ds object
Number of observations :  76 
Distance range         :  0  -  900 
AIC                    :  974.1695 
Optimisation           :  mrds (nlminb) 

Detection function:
 Hazard-rate key function 

Detection function parameters 
Scale coefficient(s): 
              estimate        se
(Intercept)  7.4142067 0.9788237
ss          -0.9677671 0.3093188

Shape coefficient(s):  
             estimate        se
(Intercept) 0.1165528 0.2173693

                       Estimate           SE        CV
Average p             0.1571083   0.05817749 0.3703019
N in covered region 483.7427660 187.54592759 0.3876976

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 0.0340516 p-value = 0.96104

*** Results for species MN ***

Summary for ds object
Number of observations :  198 
Distance range         :  0  -  2400 
AIC                    :  3044.173 
Optimisation           :  mrds (nlminb) 

Detection function:
 Half-normal key function 

Detection function parameters 
Scale coefficient(s): 
                estimate        se
(Intercept)    7.2680451 0.2949192
swell         -0.2663486 0.1016373
f_obs_heightB  0.3645463 0.2421155
f_obs_heightC  0.5825129 0.2583406
ss             0.1967999 0.1089339

                       Estimate          SE         CV
Average p             0.6228301  0.03936218 0.06319890
N in covered region 317.9037304 24.76184614 0.07789102

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 0.0981513 p-value = 0.594014

*** Results for species PD ***

Summary for ds object
Number of observations :  71 
Distance range         :  0  -  900 
AIC                    :  911.1524 
Optimisation           :  mrds (nlminb) 

Detection function:
 Hazard-rate key function 

Detection function parameters 
Scale coefficient(s): 
               estimate       se
(Intercept)    5.269611 0.661910
f_obs_heightB -2.657579 1.094798
f_obs_heightC -3.638981 1.131979

Shape coefficient(s):  
            estimate        se
(Intercept)        0 0.1641198

                       Estimate           SE        CV
Average p             0.0826893   0.03646949 0.4410425
N in covered region 858.6359047 392.91291707 0.4576013

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 0.108454 p-value = 0.544998

*** Results for species TT ***

Summary for ds object
Number of observations :  58 
Distance range         :  0  -  1200 
AIC                    :  778.8389 
Optimisation           :  mrds (nlminb) 

Detection function:
 Hazard-rate key function 

Detection function parameters 
Scale coefficient(s): 
              estimate        se
(Intercept)   4.359575 0.5885562
f_obs_heightB 1.155616 0.6384485
f_obs_heightC 1.529357 0.9459669

Shape coefficient(s):  
             estimate        se
(Intercept) 0.3391193 0.2292214

                       Estimate          SE        CV
Average p             0.2282558  0.06600979 0.2891922
N in covered region 254.1009336 79.58804189 0.3132143

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 0.02303 p-value = 0.993195

*** Results for species UC ***

Summary for ds object
Number of observations :  51 
Distance range         :  0  -  2400 
AIC                    :  776.6319 
Optimisation           :  mrds (nlminb) 

Detection function:
 Hazard-rate key function 

Detection function parameters 
Scale coefficient(s): 
              estimate        se
(Intercept)  7.6883855 1.0136627
swell       -0.3627924 0.2474246

Shape coefficient(s):  
             estimate        se
(Intercept) 0.2924517 0.3591576

                      Estimate        SE        CV
Average p             0.380769  0.115398 0.3030657
N in covered region 133.939459 43.462335 0.3244924

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 0.079092 p-value = 0.697123

*** Results for species UD ***

Summary for ds object
Number of observations :  93 
Distance range         :  0  -  1200 
AIC                    :  1304.275 
Optimisation           :  mrds (nlminb) 

Detection function:
 Hazard-rate key function 

Detection function parameters 
Scale coefficient(s): 
                estimate        se
(Intercept)    9.0337169 1.2966690
ss            -0.6853373 0.3517550
f_obs_heightB  0.2185976 0.8056287
f_obs_heightC -1.6477328 0.7820133

Shape coefficient(s):  
             estimate        se
(Intercept) 0.5442325 0.3993671

                       Estimate         SE        CV
Average p             0.5305081  0.1545731 0.2913680
N in covered region 175.3036278 53.3326810 0.3042303

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 0.072404 p-value = 0.736895

*** Results for species ULW ***

Summary for ds object
Number of observations :  422 
Distance range         :  0  -  2400 
AIC                    :  6523.225 
Optimisation           :  mrds (nlminb) 

Detection function:
 Hazard-rate key function 

Detection function parameters 
Scale coefficient(s): 
              estimate         se
(Intercept)  8.8961879 0.52986974
swell       -0.1053912 0.05803776
ss          -0.2555650 0.13023059

Shape coefficient(s):  
             estimate        se
(Intercept) 0.6759718 0.2398874

                      Estimate          SE         CV
Average p             0.738208  0.04226559 0.05725432
N in covered region 571.654593 36.05125675 0.06306475

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 0.0382295 p-value = 0.942055

5.4 Species narratives and model selection

In this section I go through species-by-species commenting on the fits and selecting which model to use for inference, based on the outputs shown above and some other exporations where appropriate.

#Create a table to store summary of results, and fill it assuming the AIC-best model is the chosen model
results_table <- truncation[, c("species", "w", "n_after")]
for(i in 1:n_species){
  results_table[i, "AIC_ranking"] <- 1
  results_table[i, "Delta_AIC"] <- 0
  results_table[i, c("key", "adjustment", "nadj", "covar")] <- det_func_tables[[i]][1, c("key", "adjustment", "nadj", "covar")]
}
#This variable is used later when selecting different models
key_model_columns <- c("key", "adjustment", "nadj", "covar", "Delta_AIC")

5.4.1 BM - Blue whale

Uniform + cosine order 1 was the AIC-best fit; 6 other models are within 2 AIC points including half normal and unif + cos (1, 2); also sea state and size in some. Nevertheless, going with the AIC-best model, which looks fine (\(g(w) \approx 0.1\) and CvM far from significant). Probability of detection estimated to be 0.5 at about 1200m. Overage p (to 2400m) is 0.55. Very similar fit to that of Campbell et al. (2015).

5.4.2 BP - Fin whale

Half-normal with 3 covariates (ss, size and f-obs-height) is AIC-best fit, with swell as a 4th covariate being 0.58 AIC points lower and a 2-covariate model without obs height being 1.78 AIC points lower. Fitted covariate values look okay - detectability goes down with increasing sea state, up with size, and up for f-obs-height C – but it is estimate to be lower with f-obs-height B which is not intuitive (although the SE is close to the estimate so not strongly significant). Overall fit looks fine, CvM test looks good. Average p (to 2400m) 0.49. Overall fit similar to Campbell et al. (2015) although in this case covars are selected – n is 231 this time compared with 166 in the Campbell analysis so that might be part of it (although sample size also larger for blue whales and no covariates chosen).

5.4.3 DC - Long-beaked common dolphin

Half-normal with size selected (and detectability goes up as size goes up); hn alone is second best with delta AIC 1.59. (As noted previously, there are some really large school sizes in the data for the common dolphin species, particularly DD, the next species after this, although there are five times more DD sightings than DC.) CvM is fine. Estimated p (to 900m) is 0.55.

This species was not considered by Campell et al. - probably because there were not enough sightings at that time.

5.4.4 DD - Short-beaked common dolphin

The AIC-best model is a hazard rate with covariates sea state (large negative effect on detectability) and size (positive effect, about half the magnitude of that for long-beaked common dolphin). As noted previously, there is a big spike in detections at smaller distances, and this results in a spiked detection function, although the shape coefficient is not 0 (it is 0.1428 so 1.15 on the back-transformed scale). The detection functon has a long tail, which corresponds to higher detectability at low ss and large school size. The average p to 900m is 0.19.

Other models within 2 AIC points either have swell as an additional covariate or drop size. The nearest half-normal model is 26 AIC points behind – a half-normal with cosine(2, 3).

In Campbell et al. (2015), we fitted separate detection functions to school sizes greater or less than 20 individuals. Below we show histograms of the different group sizes:

p <- obs_table %>% 
  filter(species == "DD" & distance < 900) %>%
  mutate(group_bin = ifelse(size > 20, "size > 20", "size <= 20")) %>% 
  ggplot( aes(x = distance) ) + geom_histogram(binwidth = 50) +
    facet_wrap(~group_bin)
p

It does look like the close distances are restricted to the smaller groups.

This issue did not appear to be present in the data analyzed by Campbell et al. – for example it is not evident in the detection function plots for this species in Figure 9 of that paper. That paper used data from 2013 and earlier, so here’s a breakup of the data to check, going out to 600m and with a 50m bin width as with the paper:

p <- obs_table %>% 
  filter(species == "DD" & distance < 600) %>%
  mutate(year_bin = ifelse(year > 2013, "year > 2013", "year <= 2013")) %>% 
  mutate(group_bin = ifelse(size > 20, "size > 20", "size <= 20")) %>% 
  ggplot( aes(x = distance) ) + geom_histogram(binwidth = 50, boundary = 50) +
    facet_wrap(~year_bin + group_bin)
p

So, it only appears to have started happening after 2013. Here is a breakdown by year:

p <- obs_table %>% 
  filter(species == "DD" & distance < 600) %>%
  ggplot( aes(x = distance) ) + geom_histogram(binwidth = 50, boundary = 50) +
    facet_wrap(~year)
p

And here is a breakdown by cruise, only for the 2014 and later cruises:

p <- obs_table %>% 
  filter(species == "DD" & distance < 400 & year > 2013) %>%
  ggplot( aes(x = distance) ) + geom_histogram(binwidth = 50, boundary = 50) +
    facet_wrap(~Region.Label)
p

2015-01 seems one cruise with a particularly large number of small distances (and 2015 in general); also 2017-01 and 2018-04. It would be worth digging into these some more; as an example here are all the sightings $<$100m for 2015 from the obs_table (a proper investigation would need going back to the distance and angle data):

tmp <- obs_table %>% 
  filter(species == "DD" & distance < 100 & year == 2015) %>%
  select(Region.Label, date_time, sighting_no, distance, size)
print(tmp, n = 100, width = 100)

# A tibble: 17 × 5
   Region.Label date_time           sighting_no distance  size
   <chr>        <dttm>              <chr>          <dbl> <dbl>
 1 CC2015-01    2015-01-17 13:59:00 29             15.5     41
 2 CC2015-01    2015-01-19 11:25:00 30             17.3     43
 3 CC2015-01    2015-01-20 13:14:00 31             26.8     98
 4 CC2015-01    2015-01-23 12:13:00 43              1.31    17
 5 CC2015-01    2015-01-25 07:01:00 47              2.59     6
 6 CC2015-01    2015-01-25 07:40:00 48             94.1     32
 7 CC2015-01    2015-01-29 09:35:00 88             18.8      5
 8 CC2015-04    2015-04-09 08:52:00 21             85.1     14
 9 CC2015-04    2015-04-09 09:36:00 22              3.83    12
10 CC2015-04    2015-04-09 10:25:00 23              5.18    40
11 CC2015-04    2015-04-16 12:38:00 59             58.8     31
12 CC2015-04    2015-04-17 17:15:00 64              1.38     3
13 CC2015-07    2015-07-14 07:28:00 27              0        3
14 CC2015-07    2015-07-14 18:21:00 28             70.8      7
15 CC2015-07    2015-07-14 19:30:00 30             16.3     14
16 CC2015-07    2015-07-22 11:18:00 125            69.4     30
17 CC2015-07    2015-07-22 14:42:00 134             0       17

The small-distance sightings are sometimes successive sighting numbers, but are not close together in time, so are separate groups.

I’ll stop looking into this for now, but What to do about it depends on what we think is causing the spike at small distances, that only appeared in later surveys.

Note

For discussion with Simone and Michaela.

For now I leave the AIC-best model as the selected model.

5.4.5 DSP - Unidentified common dolphin

The AIC-best model is like that for DD: hazard rate with a strong negative effect of swell and positive effect of size that is similar to DC - i.e., about twice as strong as DD. Together, these covariates have a very strong effect on detection probability as can be seen by the scatter of estimated detection probabilities on the detection functon plot. One notable thing is that the shape parameter estimate is 0 on the log scale, so there is effectively no shoulder in the detection function.

Alternative models within 2 AIC points have alternative covariates. There is a model (6th from top in the detection function model table) 2.84 AIC points worse that is hn with ss and size as covariates. To explore this, we plot and summarize it here:

  i <- 5
  det_table_line <- 6
  dat <- obs_table[obs_table$species == truncation$species[i], ]
  tmp <- ds(dat, truncation$w[i], key = det_func_tables[[i]]$key[det_table_line], 
            adjustment = det_func_tables[[i]]$adjustment[det_table_line],
            nadj = det_func_tables[[i]]$nadj[det_table_line], 
            region_table = region_table, sample_table = sample_table,
            formula = eval(parse(text = paste("~", det_func_tables[[i]]$covar[det_table_line]))), 
            convert_units = convert_units, quiet = TRUE)

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 2871.271

  print(summary(tmp$ddf))

Summary for ds object
Number of observations :  214 
Distance range         :  0  -  900 
AIC                    :  2871.271 
Optimisation           :  mrds (nlminb) 

Detection function:
 Half-normal key function 

Detection function parameters 
Scale coefficient(s): 
                estimate          se
(Intercept)  6.946093218 0.338685574
ss          -0.307325986 0.084055925
size         0.009670237 0.005421469

                       Estimate          SE         CV
Average p             0.6610818  0.03883328 0.05874202
N in covered region 323.7118305 23.57809073 0.07283667

  plot(tmp)

This smooths through the spike and would be a more reasonable model if there was, for example, bow riding behavior. The detection probability at 900m is around 0.4, which is likely implausibly high – probably the pattern of distances is influenced by a non-uniform distribution of unidentified groups: one may be more likely to identify closer groups and less likely to identify farther away ones, so the density of unidentified groups increases with distance from the line. This leads to “too flat” a pattern of detections and hence with a naive analysis an overestimation of \(p\) and an underestimation of \(N\).

Plotting the distance data by year (before or after 2013) and size we see some evidence for a spike both before and after that year. Note that this “species” was not analyzed in the Campbell et al. paper.

p <- obs_table %>% 
  filter(species == "DSP" & distance < 900) %>%
  mutate(year_bin = ifelse(year > 2013, "year > 2013", "year <= 2013")) %>% 
  mutate(group_bin = ifelse(size > 20, "size > 20", "size <= 20")) %>% 
  ggplot( aes(x = distance) ) + geom_histogram(binwidth = 50, boundary = 50) +
    facet_wrap(~year_bin + group_bin)
p

We leave that AIC-best model as the selected model for now.

Note

For discussion with Simone and Michaela.

5.4.6 ER - grey whale

The AIC-best model is half normal with 1 cosine adjustment; the mean \(p\) out to 2400m is 0.47 and the fit looks good (CvM p-value 0.99). (7 models are within 2 AIC points of the best model, but there seems little need to look at them closely.)

5.4.7 GG - Risso’s dolphin

The AIC-best model is a hazard rate with size and sea state as covariates. The estimated average \(p\) out to 1200m is 0.71, due to the wide shoulder – the detection prob does drop down to around 0.1 to 0.2 by w, so looks like truncation distance is fine. Detection is estimated to be certain out to 1200m for low sea states and higher group sizes, but declines rapidly with worse sea states. I’m not sure how believable this is?

There is a second model with almost the same AIC (to 2 dp): half-normal with sea state as covariate. Here’s what that looks like:

  i <- 7
  det_table_line <- 2
  dat <- obs_table[obs_table$species == truncation$species[i], ]
  tmp <- ds(dat, truncation$w[i], key = det_func_tables[[i]]$key[det_table_line], 
            adjustment = det_func_tables[[i]]$adjustment[det_table_line],
            nadj = det_func_tables[[i]]$nadj[det_table_line], 
            region_table = region_table, sample_table = sample_table,
            formula = eval(parse(text = paste("~", det_func_tables[[i]]$covar[det_table_line]))), 
            convert_units = convert_units, quiet = TRUE)

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 836.89

  print(summary(tmp$ddf))

Summary for ds object
Number of observations :  60 
Distance range         :  0  -  1200 
AIC                    :  836.89 
Optimisation           :  mrds (nlminb) 

Detection function:
 Half-normal key function 

Detection function parameters 
Scale coefficient(s): 
              estimate        se
(Intercept)  7.7328822 0.7733309
ss          -0.4715226 0.2295874

                       Estimate          SE        CV
Average p             0.5998955  0.08447765 0.1408206
N in covered region 100.0174134 16.60175958 0.1659887

  plot(tmp)

  gof_ds(tmp)

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 0.131286 p-value = 0.451995

Average detection prob is 0.6, so a bit lower. Detection probability is not certain out to 1200m even in the lowest sea state - I wonder if this is more believable? On the other hand, perhaps the lack of reliance on school size is less belivable? School size is mostly low, but there are some large schools:

hist(dat$size, breaks = 20)

There are two other models within 2 AIC points of the best two models above. These are half-normal with sea state and also school size or swell added.

Note the sample size (60) for this species is pretty low. It was not analyzed by Campbell it al.

For now we use the AIC-best model.

Note

Model-selection for discussion with Simone and Michaela.

5.4.8 LO - Pacific white-sided dolphin

The AIC-best model is hazard rate with sea state as the covariate. This has an unrealistic spike at small distances. Campbell et al. noted that this species is prone to bow-ride and also showed an unrealistic spike in their analyses, so they restricted model choice to half-normal key functions. For us, the AIC-best half-normal model is the 4th overall, with an AIC difference of 2.75 from the best hazard-rate model. It also has sea state as a covariate. We show it below:

  i <- 8
  det_table_line <- 4
  dat <- obs_table[obs_table$species == truncation$species[i], ]
  tmp <- ds(dat, truncation$w[i], key = det_func_tables[[i]]$key[det_table_line], 
            adjustment = det_func_tables[[i]]$adjustment[det_table_line],
            nadj = det_func_tables[[i]]$nadj[det_table_line], 
            region_table = region_table, sample_table = sample_table,
            formula = eval(parse(text = paste("~", det_func_tables[[i]]$covar[det_table_line]))), 
            convert_units = convert_units, quiet = TRUE)

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 976.924

  print(summary(tmp$ddf))

Summary for ds object
Number of observations :  76 
Distance range         :  0  -  900 
AIC                    :  976.9236 
Optimisation           :  mrds (nlminb) 

Detection function:
 Half-normal key function 

Detection function parameters 
Scale coefficient(s): 
              estimate        se
(Intercept)  7.7885614 0.6362711
ss          -0.5917052 0.1624722

                       Estimate        SE        CV
Average p             0.3947979  0.044739 0.1133213
N in covered region 192.5035604 29.028521 0.1507947

  plot(tmp)

  gof_ds(tmp)

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 0.64724 p-value = 0.0171269

The fit overall looks similar to that of Campbell et al., although as expected the CvM goodness-of-fit is poor. Sea state is estimated to have a negative effect, as we wold expect.

We select this half-normal model for this species:

fits[[i]] <- tmp
results_table[i, "AIC_ranking"] <- det_table_line
results_table[i, key_model_columns] <- 
  det_func_tables[[i]][det_table_line, key_model_columns]

Note that we could truncate at 700m and loose only 2 detections, estimated detection prob at this distance would be about 0.2. No pressing reason to do this however, unless it’s well warranted for other species.

5.4.9 MN - Humpback whale

The AIC-best model is a half-normal with swell, observer height and sea state as covariates. The fit looks good (CvM fine, and covars reasonable) except that detectability is estimated to increase with sea state – perhaps this is reasonable once swell is taken into account?

Estimated average detection prob is 0.2 at 2400m, so could potentially truncate a bit further out – but probably best to keep all large whales at the same value.

There are 8 models in total within 2 AIC points of the best one, with the second best (uniform + 1 cos) being only 0.36 points worse. Given the desire to predict by year, probably keeping with the AIC-best model (with the additional covars) is better? I do note, however, that the uniform + 1 cos is the same model as was selected for blue whale. Half-normal with 3 covariates was selected for fin whale (although covars different).

Going with the AIC-best model, average \(p\) to 2400m was 0.62 (sample size 198), so a bit higher than blue and fin whale.

5.4.10 PD - Dall’s porpoise

The AIC-best model is hazard rate with platform height as covariate - with higher platforms estimated to have lower detectability, which seems counter-intuitive but it could be because they have worse bow-riding behaviour possibly? The fitted function has a huge spike, and this is another of the species that Campbell et al. fitted with a half-normal rather than hazard rate.

The AIC-best half-normal nodel is 10th from top, with sea state as covariate. The fit is as follows:

  i <- 10
  det_table_line <- 10
  dat <- obs_table[obs_table$species == truncation$species[i], ]
  tmp <- ds(dat, truncation$w[i], key = det_func_tables[[i]]$key[det_table_line], 
            adjustment = det_func_tables[[i]]$adjustment[det_table_line],
            nadj = det_func_tables[[i]]$nadj[det_table_line], 
            region_table = region_table, sample_table = sample_table,
            formula = eval(parse(text = paste("~", det_func_tables[[i]]$covar[det_table_line]))), 
            convert_units = convert_units, quiet = TRUE)

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 925.357

  print(summary(tmp$ddf))

Summary for ds object
Number of observations :  71 
Distance range         :  0  -  900 
AIC                    :  925.3569 
Optimisation           :  mrds (nlminb) 

Detection function:
 Half-normal key function 

Detection function parameters 
Scale coefficient(s): 
              estimate         se
(Intercept)  6.3711408 0.29892402
ss          -0.1666811 0.07794922

                       Estimate         SE         CV
Average p             0.4550069  0.0376224 0.08268534
N in covered region 156.0416026 18.9046862 0.12115158

  plot(tmp)

  gof_ds(tmp)

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 0.464543 p-value = 0.0490671

This fit looks okay - sea state estimated to have a small negative effect, CvM goodness of fit is marginally significant, but this is expected given this is not the AIC-best model and we’re fitting it for a reason (to deal with bow riding). Estimated detection prob is fairly low at truncation distance so may be better to revert to a 700m truncation distance (as was used by Campbell et al). But OK to leave for now.

So, we select the HN model:

fits[[i]] <- tmp
results_table[i, "AIC_ranking"] <- det_table_line
results_table[i, key_model_columns] <- 
  det_func_tables[[i]][det_table_line, key_model_columns]

5.4.11 TT - Bottlenose dolphin

The AIC-best model is a hazard rate with platform height as a covariate; the values of platform height look sensible. The fit has a fair spike, so another possible candidate for a half-normal fit. The AIC-best half-normal model is 5th, 1.82 AIC points worse - it’s a hn with 1 cosine adjusment (no covariates):

  i <- 11
  det_table_line <- 5
  dat <- obs_table[obs_table$species == truncation$species[i], ]
  #No covariate model
  tmp <- ds(dat, truncation$w[i], key = det_func_tables[[i]]$key[det_table_line], 
            adjustment = det_func_tables[[i]]$adjustment[det_table_line],
            nadj = det_func_tables[[i]]$nadj[det_table_line], 
            region_table = region_table, sample_table = sample_table, 
            convert_units = convert_units, quiet = TRUE)

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2) adjustments

AIC= 780.661

  print(summary(tmp$ddf))

Summary for ds object
Number of observations :  58 
Distance range         :  0  -  1200 
AIC                    :  780.6615 
Optimisation           :  mrds (slsqp) 

Detection function:
 Half-normal key function with cosine adjustment term of order 2 

Detection function parameters 
Scale coefficient(s): 
            estimate        se
(Intercept) 6.184463 0.1184894

Adjustment term coefficient(s):  
              estimate        se
cos, order 2 0.5456947 0.1816031


Strict monotonicity constraints were enforced.
                       Estimate          SE        CV
Average p             0.3291332  0.03779182 0.1148223
N in covered region 176.2204377 27.72369318 0.1573239

Strict monotonicity constraints were enforced.

  plot(tmp)

  gof_ds(tmp)

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 0.0501764 p-value = 0.875209

Looks like a sensible fit to the data, so select it for inference.

fits[[i]] <- tmp
results_table[i, "AIC_ranking"] <- det_table_line
results_table[i, key_model_columns] <- 
  det_func_tables[[i]][det_table_line, key_model_columns]

This species was not fitted by Campbell et al. so no comnparison there.

Note

Check OK with Simone and Michaela.

5.4.12 UC - Unidentified cetacean

The AIC-best model is hazard rate with swell as a covariate (detecton prob goes down as swell goes up); the fit according to CvM is fine.

7 models are within 2 AIC points of the best (not surprising with the small sample size of 51). The next best is hn with size as covariate.

For now, go with the AIC best model. Estimated detection prob out to 2400m is 0.38, which is quite low – it would probably be rather higher with the hn model.

5.4.13 UD - Unidentified dolphin

The AIC-best model is a hazard rate (with a small shoulder) and covariates sea state and obs height. Detection prob decreases with increasing sea state and for the middle platform but is then estimated to decrease at the highest platform, which seems counter-intuitive.

We note that these UID “species” are tricky to model due to a likely non-uniform distribution of animals wrt distance from the transect line.

There are 4 models with an AIC difference < 2 from the AIC-best model. Hn + ss + obs height + swell is the next best model with an AIC difference of 0.34.

  i <- 13
  det_table_line <- 2
  dat <- obs_table[obs_table$species == truncation$species[i], ]
  tmp <- ds(dat, truncation$w[i], key = det_func_tables[[i]]$key[det_table_line], 
            adjustment = det_func_tables[[i]]$adjustment[det_table_line],
            nadj = det_func_tables[[i]]$nadj[det_table_line], 
            region_table = region_table, sample_table = sample_table,
            formula = eval(parse(text = paste("~", det_func_tables[[i]]$covar[det_table_line]))), 
            convert_units = convert_units, quiet = TRUE)

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 1304.616

  print(summary(tmp$ddf))

Summary for ds object
Number of observations :  93 
Distance range         :  0  -  1200 
AIC                    :  1304.616 
Optimisation           :  mrds (nlminb) 

Detection function:
 Half-normal key function 

Detection function parameters 
Scale coefficient(s): 
                 estimate        se
(Intercept)    7.16568944 0.6017587
ss            -0.27501434 0.1474605
f_obs_heightB -0.04042542 0.6556494
f_obs_heightC -1.17718305 0.5382958
swell          0.19550275 0.1359591

                       Estimate          SE        CV
Average p             0.6380437  0.07684412 0.1204371
N in covered region 145.7580472 20.26289695 0.1390173

  plot(tmp)

  gof_ds(tmp)

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 0.055712 p-value = 0.84095

This has the same issue with the largest platform height, plus a counter-intuitive effect of swell. So, for now, stick with the AIC-best model.

5.4.14 ULW - Unidentified large whale

This has a large sample size (422). The AIC-best model is hazard rate with swell and sea state as covariates, both decreasing detecability. Detection prob is esitmated to be certain out to 500m and increasingly variable after that. There are 7 models within 2 AIC points of the best model; the first with a hn detection function is 4th, with sea state, swell and obs height:

  i <- 14
  det_table_line <- 4
  dat <- obs_table[obs_table$species == truncation$species[i], ]
  tmp <- ds(dat, truncation$w[i], key = det_func_tables[[i]]$key[det_table_line], 
            adjustment = det_func_tables[[i]]$adjustment[det_table_line],
            nadj = det_func_tables[[i]]$nadj[det_table_line], 
            region_table = region_table, sample_table = sample_table,
            formula = eval(parse(text = paste("~", det_func_tables[[i]]$covar[det_table_line]))), 
            convert_units = convert_units, quiet = TRUE)

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 6524.561

  print(summary(tmp$ddf))

Summary for ds object
Number of observations :  422 
Distance range         :  0  -  2400 
AIC                    :  6524.561 
Optimisation           :  mrds (nlminb) 

Detection function:
 Half-normal key function 

Detection function parameters 
Scale coefficient(s): 
                estimate         se
(Intercept)    8.5882515 0.52593046
ss            -0.2104529 0.11421732
swell         -0.1014254 0.04573649
f_obs_heightB  0.3548913 0.19972165
f_obs_heightC  0.3059925 0.19244298

                       Estimate          SE         CV
Average p             0.7206665  0.03192934 0.04430529
N in covered region 585.5690167 30.41241101 0.05193651

  plot(tmp)

  gof_ds(tmp)

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 0.0620967 p-value = 0.800778

This has a very similar average \(p\) of 0.72 to 2400m, and the covariate estimates look sensible (obs height C is about the same as obs height B, but perhaps not too surprising for large whales to 2400m). This model may be preferred, but for now will stick with the AIC-best model.

5.4.15 Summary and final detection function results table

Here is a summary of the fitted models:

for(i in 1:n_species){
  tmp <- summary(fits[[i]]$ddf)
  results_table[i, "p"] <- tmp$average.p
  results_table[i, "cv_p"] <- as.numeric(tmp$average.p.se / tmp$average.p)
}
results_table %>% arrange(desc(w), species) %>% 
  kbl %>% kable_classic(full_width = FALSE)

species	w	n_after	AIC_ranking	Delta_AIC	key	adjustment	nadj	covar	p	cv_p
BM	2400	143	1	0.000000	unif	cos	1		0.5498343	0.0397623
BP	2400	231	1	0.000000	hn	cos	0	ss + size + f_obs_height	0.4913724	0.0557339
ER	2400	62	1	0.000000	hn	cos	1		0.4767994	0.1538231
MN	2400	198	1	0.000000	hn	cos	0	swell + f_obs_height + ss	0.6228301	0.0631989
UC	2400	51	1	0.000000	hr	poly	0	swell	0.3807690	0.3030657
ULW	2400	422	1	0.000000	hr	poly	0	swell + ss	0.7382080	0.0572543
GG	1200	60	1	0.000000	hr	poly	0	size + ss	0.7052541	0.0789592
TT	1200	58	5	1.822648	hn	cos	1		0.3291332	0.1148223
UD	1200	93	1	0.000000	hr	poly	0	ss + f_obs_height	0.5305081	0.2913680
DC	900	77	1	0.000000	hn	cos	0	size	0.5540501	0.0927915
DD	900	497	1	0.000000	hr	poly	0	ss + size	0.1904508	0.1131744
DSP	900	214	1	0.000000	hr	poly	0	swell + size	0.3356156	0.2862127
LO	900	76	4	2.754054	hn	cos	0	ss	0.3947979	0.1133213
PD	900	71	10	14.204571	hn	cos	0	ss	0.4550069	0.0826853

5.5 Detection function modelling combining species

Here we try modelling all the larger species, with species as a covariate.

  w <- 2400
  species <- truncation$species[truncation$w == w]
  dat <- obs_table[obs_table$species %in% species, ]
  tmp <- fit_det_func(dat, w = w,
    key = c("unif", "hn"),
    adjustments = c("cos", "cos"),
    max_adjustments = c(3, 2),
    covariates = c("ss", "swell", "size", "f_obs_height", "species"),
    covariates_and_adjustments = TRUE)

Warning: Unknown or uninitialised column: `distbegin`.

Fitting uniform key function

AIC= 17232.058

Warning: Unknown or uninitialised column: `distbegin`.

Fitting uniform key function with cosine(1) adjustments

AIC= 16998.102

Warning: Unknown or uninitialised column: `distbegin`.

Fitting uniform key function with cosine(1,2) adjustments

AIC= 16994.074

Warning: Unknown or uninitialised column: `distbegin`.

Fitting uniform key function with cosine(1,2,3) adjustments

AIC= 16995.233

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 17007.285

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 17000.359

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 16994.309

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 17008.445

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 17007.377

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 16987.11

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 16972.688

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 16966.31

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 16985.85

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 16990.002

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 16966.217

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 16963.118

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 16965.048

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 16962.244

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 16961.564

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 16960.53

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2) adjustments

AIC= 17001.059

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2) adjustments

AIC= 16994.832

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2) adjustments

Warning in optimx.check(par, optcfg$ufn, optcfg$ugr, optcfg$uhess, lower, : Parameters or bounds appear to have different scalings.
  This can cause poor performance in optimization. 
  It is important for derivative free methods like BOBYQA, UOBYQA, NEWUOA.

AIC= 16989.276

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2) adjustments

AIC= 17002.312

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2) adjustments

AIC= 17001.444

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2) adjustments

AIC= 16983.041

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2) adjustments

AIC= 16969.68

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2) adjustments

AIC= 16963.772

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2) adjustments

AIC= 16982.053

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2) adjustments

AIC= 16985.97

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2) adjustments

AIC= 16963.773

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2) adjustments

AIC= 16960.999

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2) adjustments

Warning in optimx.check(par, optcfg$ufn, optcfg$ugr, optcfg$uhess, lower, : Parameters or bounds appear to have different scalings.
  This can cause poor performance in optimization. 
  It is important for derivative free methods like BOBYQA, UOBYQA, NEWUOA.

AIC= 16962.867

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2) adjustments

AIC= 16960.301

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2) adjustments

Warning in optimx.check(par, optcfg$ufn, optcfg$ugr, optcfg$uhess, lower, : Parameters or bounds appear to have different scalings.
  This can cause poor performance in optimization. 
  It is important for derivative free methods like BOBYQA, UOBYQA, NEWUOA.

AIC= 16959.826

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2) adjustments

AIC= 16959.002

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2,3) adjustments

AIC= 16998.454

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2,3) adjustments

AIC= 16992.439

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2,3) adjustments

AIC= 16986.997

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2,3) adjustments

AIC= 16999.74

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2,3) adjustments

AIC= 16998.928

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2,3) adjustments

AIC= 16981.101

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2,3) adjustments

AIC= 16967.986

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2,3) adjustments

AIC= 16962.188

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2,3) adjustments

AIC= 16980.197

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2,3) adjustments

AIC= 16984.038

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2,3) adjustments

AIC= 16962.213

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2,3) adjustments

AIC= 16959.511

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2,3) adjustments

Warning in optimx.check(par, optcfg$ufn, optcfg$ugr, optcfg$uhess, lower, : Parameters or bounds appear to have different scalings.
  This can cause poor performance in optimization. 
  It is important for derivative free methods like BOBYQA, UOBYQA, NEWUOA.

AIC= 16961.341

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2,3) adjustments

AIC= 16958.854

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2,3) adjustments

Warning in optimx.check(par, optcfg$ufn, optcfg$ugr, optcfg$uhess, lower, : Parameters or bounds appear to have different scalings.
  This can cause poor performance in optimization. 
  It is important for derivative free methods like BOBYQA, UOBYQA, NEWUOA.

AIC= 16958.391

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2,3) adjustments

AIC= 16957.611

  det_table_line <- 1
  tmp2 <- ds(dat, w, key = tmp$key[det_table_line], 
            adjustment = tmp$adjustment[det_table_line],
            nadj = tmp$nadj[det_table_line], 
            region_table = region_table, sample_table = sample_table,
            formula = eval(parse(text = paste("~", tmp$covar[det_table_line]))), 
            convert_units = convert_units, quiet = TRUE)

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2,3) adjustments
AIC= 16957.611

  print(summary(tmp2$ddf))

Summary for ds object
Number of observations :  1107 
Distance range         :  0  -  2400 
AIC                    :  16957.61 
Optimisation           :  mrds (nlminb) 

Detection function:
 Half-normal key function with cosine adjustment terms of order 2, 3 

Detection function parameters 
Scale coefficient(s): 
                 estimate         se
(Intercept)    7.41451900 0.15264325
swell         -0.09922037 0.02599868
size           0.06970337 0.03940414
f_obs_heightB  0.08054105 0.08384607
f_obs_heightC  0.22283966 0.10361601
speciesBP     -0.03770087 0.10004908
speciesER     -0.05649214 0.16500687
speciesMN      0.12753813 0.11872773
speciesUC      0.14271969 0.14528265
speciesULW     0.49768911 0.11181798
ss            -0.05904602 0.03219374

Adjustment term coefficient(s):  
                estimate         se
cos, order 2  0.10708276 0.04458476
cos, order 3 -0.08136436 0.04404440

                        Estimate           SE         CV
Average p              0.5997285   0.03267451 0.05448217
N in covered region 1845.8351654 106.89204383 0.05790985

  plot(tmp2)

  par(mfrow = c(2,3))
  for(i in 1:6){
    plot(tmp2, breaks = c(0, 400, 800, 1200, 1600, 2000, 2400), subset = species == species[i], main = species[i])
  }

  par(mfrow = c(1, 1))
  gof_ds(tmp2)

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 0.082335 p-value = 0.678462

All of the parameter estimates look sensible to me.

Could compare by summing the AICs of the by-species models, although this is bound to be a bit higher.

Here we try the smaller ones, using 900m as the truncation distance.

  w <- 900
  species <- truncation$species[truncation$w %in% c(900, 1200)]
  dat <- obs_table[obs_table$species %in% species, ]
  tmp <- fit_det_func(dat, w = w,
    key = c("unif", "hn"),
    adjustments = c("cos", "cos"),
    max_adjustments = c(3, 2),
    covariates = c("ss", "swell", "size", "f_obs_height", "species"),
    covariates_and_adjustments = FALSE)

Warning: Unknown or uninitialised column: `distbegin`.

Fitting uniform key function

AIC= 15291.783

Warning: Unknown or uninitialised column: `distbegin`.

Fitting uniform key function with cosine(1) adjustments

AIC= 14879.494

Warning: Unknown or uninitialised column: `distbegin`.

Fitting uniform key function with cosine(1,2) adjustments

AIC= 14842.173

Warning: Unknown or uninitialised column: `distbegin`.

Fitting uniform key function with cosine(1,2,3) adjustments

AIC= 14799.291

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 14873.215

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 14808.896

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 14822.783

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 14869.667

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 14875.625

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 14798.005

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 14763.281

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 14777.635

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 14792.25

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 14800.537

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 14761.771

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 14755.844

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 14764.633

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 14755.504

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 14757.594

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 14757.33

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function with cosine(2) adjustments

AIC= 14806.835

  det_table_line <- 1
  tmp2 <- ds(dat, w, key = tmp$key[det_table_line], 
            adjustment = tmp$adjustment[det_table_line],
            nadj = tmp$nadj[det_table_line], 
            region_table = region_table, sample_table = sample_table,
            formula = eval(parse(text = paste("~", tmp$covar[det_table_line]))), 
            convert_units = convert_units, quiet = TRUE)

Warning: Unknown or uninitialised column: `distbegin`.

Fitting half-normal key function

AIC= 14755.504

  print(summary(tmp2$ddf))

Summary for ds object
Number of observations :  1124 
Distance range         :  0  -  900 
AIC                    :  14755.5 
Optimisation           :  mrds (nlminb) 

Detection function:
 Half-normal key function 

Detection function parameters 
Scale coefficient(s): 
                estimate          se
(Intercept)  6.388419292 0.140365456
ss          -0.134450554 0.025753825
size         0.001296392 0.000438673
speciesDD   -0.129702953 0.117621633
speciesDSP   0.316731526 0.144130083
speciesGG    0.527565678 0.320553146
speciesLO   -0.115190699 0.144631787
speciesPD    0.007801440 0.145544525
speciesTT   -0.050740281 0.154620526
speciesUD    0.552222531 0.228395588
swell       -0.029595406 0.016391012

                        Estimate          SE         CV
Average p              0.5031611  0.01096696 0.02179611
N in covered region 2233.8769514 69.09085641 0.03092868

  plot(tmp2)

  par(mfrow = c(3,3))
  for(i in 1:length(species)){
    plot(tmp2, subset = (species == species[i]), main = species[i])
  }
  par(mfrow = c(1, 1))

  gof_ds(tmp2)

Goodness of fit results for ddf object

Distance sampling Cramer-von Mises test (unweighted)
Test statistic = 3.82772 p-value = 8.3275e-06

Doesn’t look as good!

6 Density estimation

Just for the purposes of producing initial outputs, here are density estimates for the by-species detection functions.

for(i in 1:n_species) {
  tmp3 <- summary(fits[[i]])$dht$individuals$D  %>% 
    kbl(digits = 2, caption = paste("Species:", truncation$species[i])) %>%
    kable_classic(full_width = FALSE)
  print(tmp3)
}

Species: BM

Label	Estimate	se	cv	lcl	ucl	df
CC2004-07	1.57	0.76	0.49	0.58	4.24	13.18
CC2004-11	0.81	0.78	0.95	0.14	4.63	13.05
CC2005-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-07	3.78	2.55	0.68	1.01	14.18	13.09
CC2005-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2006-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2006-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2006-07	2.25	1.66	0.74	0.55	9.26	14.08
CC2006-10	1.47	0.85	0.58	0.46	4.74	12.11
CC2007-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2007-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2007-07	0.38	0.39	1.04	0.06	2.37	14.04
CC2007-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-03	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-08	0.71	0.49	0.68	0.19	2.72	13.09
CC2008-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-03	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-07	3.53	2.72	0.77	0.82	15.12	15.08
CC2009-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-04	0.31	0.32	1.04	0.05	1.92	14.04
CC2011-08	7.91	5.74	0.73	1.94	32.20	13.08
CC2011-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2012-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2012-03	0.00	0.00	0.00	0.00	0.00	0.00
CC2012-07	2.81	1.39	0.49	1.04	7.61	15.20
CC2012-10	0.80	0.54	0.68	0.21	3.03	13.09
CC2013-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-07	1.14	0.85	0.74	0.27	4.80	13.07
CC2013-11	1.23	1.23	1.00	0.20	7.44	13.04
CC2014-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-07	3.32	1.88	0.57	1.08	10.23	15.15
CC2014-11	0.33	0.33	1.01	0.05	2.00	13.04
CC2015-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2015-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2015-07	2.92	1.38	0.47	1.11	7.68	13.19
CC2015-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-04	1.48	1.12	0.76	0.33	6.53	11.06
CC2016-07	4.92	2.49	0.51	1.72	14.05	11.14
CC2016-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2017-01	0.74	0.76	1.03	0.12	4.72	12.04
CC2017-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2017-08	5.60	3.25	0.58	1.73	18.11	12.11
CC2017-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-06	0.63	0.62	0.97	0.11	3.68	13.04
CC2018-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-07	1.51	1.57	1.04	0.23	9.76	12.04
CC2019-11	1.77	1.40	0.79	0.37	8.33	10.05
CC2020-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2021-07	15.65	6.01	0.38	6.93	35.34	11.24
CC2021-11	0.46	0.47	1.03	0.07	3.03	10.03
CC2204SH	0.00	0.00	0.00	0.00	0.00	0.00
CC2208	5.27	2.06	0.39	2.33	11.97	12.25
CC2211SR	1.07	0.80	0.74	0.25	4.68	10.06
CC2301RL	0.00	0.00	0.00	0.00	0.00	0.00
CC2303SH	0.00	0.00	0.00	0.00	0.00	0.00
CC2311SR	0.41	0.42	1.04	0.06	2.59	13.04
Total	1.07	0.17	0.16	0.79	1.45	81.31

Species: BP

Label	Estimate	se	cv	lcl	ucl	df
CC2004-07	3.30	1.37	0.41	1.41	7.69	15.30
CC2004-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-07	2.51	1.26	0.50	0.90	6.98	13.64
CC2005-11	6.45	6.12	0.95	1.13	36.85	12.20
CC2006-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2006-04	0.65	0.67	1.03	0.11	4.03	14.22
CC2006-07	1.92	1.48	0.77	0.44	8.25	14.47
CC2006-10	2.12	1.86	0.88	0.41	10.94	12.53
CC2007-01	0.58	0.59	1.02	0.09	3.58	13.21
CC2007-04	0.72	0.74	1.03	0.11	4.73	10.16
CC2007-07	2.67	2.65	1.00	0.45	15.77	14.21
CC2007-11	1.06	1.02	0.97	0.18	6.17	12.24
CC2008-01	2.68	1.88	0.70	0.69	10.45	13.46
CC2008-03	0.63	0.65	1.04	0.09	4.32	9.23
CC2008-08	1.59	1.07	0.67	0.43	5.91	13.37
CC2008-10	1.04	0.78	0.75	0.24	4.41	13.45
CC2009-01	0.29	0.28	0.99	0.05	1.71	13.25
CC2009-03	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-07	1.53	1.50	0.98	0.27	8.70	15.87
CC2009-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-01	1.22	1.18	0.97	0.22	6.96	14.42
CC2010-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-11	0.86	0.90	1.05	0.13	5.60	12.32
CC2011-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-04	1.28	1.34	1.05	0.20	8.03	14.81
CC2011-08	12.25	6.23	0.51	4.38	34.23	14.13
CC2011-10	5.01	2.90	0.58	1.56	16.08	12.52
CC2012-02	0.90	0.87	0.97	0.12	6.55	6.10
CC2012-03	6.01	4.31	0.72	1.34	26.98	7.55
CC2012-07	1.60	1.00	0.63	0.47	5.43	15.94
CC2012-10	5.65	4.22	0.75	1.35	23.69	13.39
CC2013-01	1.03	1.06	1.04	0.16	6.46	13.35
CC2013-04	2.13	1.93	0.91	0.37	12.21	9.19
CC2013-07	5.23	2.32	0.44	2.11	12.97	14.02
CC2013-11	1.77	1.25	0.71	0.45	6.98	13.41
CC2014-02	1.19	1.39	1.16	0.11	12.69	5.09
CC2014-04	3.57	3.47	0.97	0.59	21.79	10.32
CC2014-07	8.85	6.62	0.75	2.14	36.55	15.44
CC2014-11	0.29	0.28	0.97	0.05	1.66	13.27
CC2015-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2015-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2015-07	4.24	3.01	0.71	1.07	16.77	13.34
CC2015-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-01	0.36	0.37	1.02	0.06	2.21	13.46
CC2016-04	1.06	0.79	0.74	0.25	4.51	11.90
CC2016-07	7.08	4.95	0.70	1.77	28.38	11.04
CC2016-11	0.45	0.44	0.98	0.07	2.69	11.35
CC2017-01	0.63	0.64	1.01	0.10	3.88	12.51
CC2017-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2017-08	1.66	1.20	0.72	0.41	6.73	12.84
CC2017-11	2.87	1.63	0.57	0.91	9.06	12.20
CC2018-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-04	1.06	0.78	0.73	0.26	4.37	13.25
CC2018-06	7.07	7.04	1.00	1.18	42.43	13.03
CC2018-10	4.82	3.11	0.64	1.36	17.06	14.03
CC2019-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-04	0.34	0.37	1.07	0.05	2.38	10.32
CC2019-07	0.86	0.60	0.69	0.22	3.35	12.44
CC2019-11	1.25	0.80	0.64	0.34	4.61	10.27
CC2020-01	1.52	1.18	0.77	0.35	6.64	13.57
CC2021-07	6.30	4.30	0.68	1.62	24.41	11.41
CC2021-11	0.97	1.04	1.08	0.14	6.77	10.58
CC2204SH	1.32	1.40	1.06	0.17	10.20	7.14
CC2208	3.27	2.34	0.71	0.81	13.26	12.10
CC2211SR	10.60	5.36	0.51	3.70	30.41	10.73
CC2301RL	0.00	0.00	0.00	0.00	0.00	0.00
CC2303SH	0.00	0.00	0.00	0.00	0.00	0.00
CC2311SR	3.65	1.25	0.34	1.81	7.38	16.02
Total	2.14	0.30	0.14	1.63	2.81	224.86

Species: DC

Label	Estimate	se	cv	lcl	ucl	df
CC2004-07	45.37	45.21	1.00	7.58	271.49	13.37
CC2004-11	1257.15	1305.14	1.04	198.17	7974.99	13.00
CC2005-01	71.66	77.43	1.08	10.38	494.59	11.19
CC2005-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-07	162.00	150.39	0.93	29.82	880.23	13.85
CC2005-11	515.71	387.35	0.75	122.06	2178.89	13.36
CC2006-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2006-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2006-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2006-10	17.15	18.70	1.09	2.51	117.28	12.34
CC2007-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2007-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2007-07	17.56	18.01	1.03	2.86	107.66	14.45
CC2007-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-03	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-01	14.48	14.79	1.02	2.35	89.39	13.45
CC2009-03	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-07	45.78	47.72	1.04	7.38	283.98	15.29
CC2009-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-01	119.56	90.70	0.76	28.30	505.15	14.53
CC2010-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-10	65.68	67.70	1.03	10.33	417.64	12.23
CC2012-02	391.51	387.10	0.99	53.87	2845.13	6.50
CC2012-03	29.80	27.59	0.93	4.70	188.99	7.26
CC2012-07	172.45	135.99	0.79	39.58	751.34	16.33
CC2012-10	8.80	9.52	1.08	1.32	58.56	13.43
CC2013-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-07	39.22	40.46	1.03	6.26	245.71	13.31
CC2013-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-07	29.00	30.79	1.06	4.57	184.11	15.35
CC2014-11	44.21	38.17	0.86	8.86	220.47	13.45
CC2015-01	3213.87	3064.41	0.95	565.71	18258.27	13.01
CC2015-04	416.45	439.35	1.05	63.91	2713.69	12.62
CC2015-07	14.48	14.36	0.99	2.44	86.03	13.44
CC2015-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-01	280.10	253.92	0.91	52.81	1485.77	13.37
CC2016-04	55.71	43.97	0.79	12.14	255.69	11.48
CC2016-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-11	1766.01	1876.01	1.06	260.74	11961.33	11.00
CC2017-01	56.40	60.14	1.07	8.48	375.14	12.22
CC2017-04	203.18	212.96	1.05	31.37	1316.15	12.45
CC2017-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2017-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-06	204.82	218.80	1.07	31.40	1336.18	13.69
CC2018-10	96.38	73.80	0.77	22.07	420.96	12.44
CC2019-02	219.64	249.13	1.13	17.87	2699.37	4.07
CC2019-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-07	378.84	405.81	1.07	57.42	2499.41	13.13
CC2019-11	126.36	126.71	1.00	19.91	801.91	10.45
CC2020-01	115.02	98.21	0.85	23.36	566.40	13.37
CC2021-07	210.27	229.82	1.09	30.53	1448.21	12.11
CC2021-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2204SH	0.00	0.00	0.00	0.00	0.00	0.00
CC2208	783.28	520.30	0.66	214.67	2858.01	14.27
CC2211SR	0.00	0.00	0.00	0.00	0.00	0.00
CC2301RL	389.22	348.47	0.90	67.27	2251.85	8.38
CC2303SH	1492.83	1208.97	0.81	319.04	6985.05	12.30
CC2311SR	620.53	630.08	1.02	100.90	3816.36	13.14
Total	195.58	61.98	0.32	104.78	365.05	42.44

Species: DD

Label	Estimate	se	cv	lcl	ucl	df
CC2004-07	1969.11	881.23	0.45	793.02	4889.44	15.23
CC2004-11	1721.54	809.54	0.47	666.24	4448.39	15.63
CC2005-01	2692.93	2764.56	1.03	442.00	16407.00	15.11
CC2005-04	10.99	10.79	0.98	1.87	64.51	13.41
CC2005-07	806.19	451.76	0.56	263.72	2464.56	14.46
CC2005-11	1126.31	755.58	0.67	301.87	4202.39	13.02
CC2006-02	478.04	322.24	0.67	130.54	1750.60	15.92
CC2006-04	1581.39	1538.88	0.97	283.14	8832.34	17.27
CC2006-07	1347.52	489.92	0.36	639.79	2838.16	16.56
CC2006-10	605.89	296.65	0.49	222.73	1648.24	13.09
CC2007-01	2506.74	1302.56	0.52	882.13	7123.44	14.62
CC2007-04	1494.58	1166.03	0.78	332.04	6727.46	11.85
CC2007-07	863.72	608.44	0.70	222.61	3351.26	14.65
CC2007-11	20.81	22.21	1.07	3.15	137.39	12.77
CC2008-01	152.69	103.38	0.68	40.89	570.19	13.99
CC2008-03	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-08	1054.23	907.06	0.86	219.77	5057.04	17.36
CC2008-10	688.99	509.57	0.74	166.58	2849.66	13.69
CC2009-01	400.05	334.64	0.84	83.40	1919.06	13.46
CC2009-03	112.07	98.58	0.88	21.65	580.14	12.36
CC2009-07	1803.44	1097.62	0.61	552.67	5884.87	17.32
CC2009-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-01	2624.53	2153.77	0.82	595.77	11561.80	23.45
CC2010-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-01	1096.19	1050.12	0.96	197.49	6084.33	15.63
CC2011-04	671.29	677.32	1.01	112.69	3998.80	15.14
CC2011-08	474.18	290.56	0.61	141.05	1594.04	13.82
CC2011-10	1843.01	1572.00	0.85	375.56	9044.40	13.55
CC2012-02	1436.28	1315.95	0.92	219.67	9390.79	6.46
CC2012-03	629.85	652.48	1.04	85.88	4619.25	7.49
CC2012-07	444.49	261.24	0.59	139.93	1411.99	15.82
CC2012-10	899.54	640.09	0.71	227.91	3550.35	13.99
CC2013-01	1057.64	792.67	0.75	251.45	4448.67	13.54
CC2013-04	203.15	207.01	1.02	30.63	1347.31	9.57
CC2013-07	2449.49	2132.11	0.87	519.52	11548.99	23.78
CC2013-11	1301.50	915.17	0.70	332.99	5086.99	13.60
CC2014-02	5785.58	2842.47	0.49	1862.44	17972.59	6.10
CC2014-04	2002.54	1081.48	0.54	656.13	6111.82	10.81
CC2014-07	309.15	162.64	0.53	108.47	881.12	16.09
CC2014-11	2125.98	1261.74	0.59	666.03	6786.14	16.68
CC2015-01	3640.76	2127.25	0.58	1130.70	11722.93	13.16
CC2015-04	468.92	203.43	0.43	192.00	1145.25	13.63
CC2015-07	786.67	468.36	0.60	241.42	2563.31	14.02
CC2015-11	147.89	151.17	1.02	23.87	916.10	13.25
CC2016-01	1273.97	487.85	0.38	578.88	2803.67	14.93
CC2016-04	566.56	356.81	0.63	161.43	1988.37	12.34
CC2016-07	598.82	380.84	0.64	168.50	2128.16	12.15
CC2016-11	1505.91	732.46	0.49	556.27	4076.72	12.95
CC2017-01	4980.36	2373.09	0.48	1889.03	13130.52	14.11
CC2017-04	2178.90	1333.17	0.61	646.10	7348.10	13.28
CC2017-08	286.89	173.12	0.60	85.80	959.25	12.67
CC2017-11	2936.15	1670.26	0.57	956.74	9010.81	16.20
CC2018-02	2849.60	1424.77	0.50	961.38	8446.38	8.12
CC2018-04	1378.03	836.92	0.61	412.10	4608.03	13.37
CC2018-06	1015.96	631.81	0.62	298.93	3452.91	14.38
CC2018-10	310.48	214.47	0.69	80.86	1192.23	13.38
CC2019-02	776.90	569.75	0.73	130.79	4614.90	4.24
CC2019-04	5065.72	2275.15	0.45	2016.29	12727.12	13.71
CC2019-07	217.51	167.26	0.77	49.80	950.07	12.84
CC2019-11	2204.50	1475.96	0.67	588.02	8264.71	12.38
CC2020-01	7194.89	3836.64	0.53	2488.91	20798.87	15.81
CC2021-07	9.43	9.51	1.01	1.50	59.29	11.26
CC2021-11	538.94	225.43	0.42	223.34	1300.46	11.31
CC2204SH	24.70	18.74	0.76	5.08	120.09	7.30
CC2208	353.74	352.57	1.00	58.54	2137.78	12.68
CC2211SR	4393.80	2089.06	0.48	1690.46	11420.21	16.39
CC2301RL	761.49	506.95	0.67	190.47	3044.39	8.38
CC2303SH	1081.88	1084.39	1.00	176.99	6613.24	12.41
CC2311SR	1329.37	648.48	0.49	498.71	3543.60	15.80
Total	1366.72	240.48	0.18	970.00	1925.69	662.32

Species: DSP

Label	Estimate	se	cv	lcl	ucl	df
CC2004-07	214.07	134.36	0.63	61.85	740.92	13.37
CC2004-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-01	15.91	15.96	1.00	2.57	98.60	11.67
CC2005-04	26.93	27.21	1.01	4.51	160.90	15.03
CC2005-07	78.17	58.18	0.74	18.80	324.95	13.91
CC2005-11	162.27	139.38	0.86	34.90	754.55	23.24
CC2006-02	10.78	10.59	0.98	1.87	62.26	14.90
CC2006-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2006-07	233.10	189.11	0.81	50.73	1071.05	14.03
CC2006-10	45.44	46.18	1.02	7.28	283.77	12.21
CC2007-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2007-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2007-07	154.49	104.93	0.68	41.32	577.71	14.24
CC2007-11	1254.53	1238.30	0.99	208.00	7566.48	12.00
CC2008-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-03	3.18	3.38	1.06	0.45	22.28	9.61
CC2008-08	17.35	12.71	0.73	4.28	70.31	14.74
CC2008-10	75.13	75.89	1.01	12.30	458.91	13.15
CC2009-01	46.44	27.11	0.58	14.43	149.43	13.13
CC2009-03	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-07	12.66	7.84	0.62	3.81	42.09	17.15
CC2009-11	61.48	50.18	0.82	13.02	290.25	12.33
CC2010-01	184.49	136.87	0.74	45.91	741.47	18.04
CC2010-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-11	18.94	19.04	1.01	3.07	116.75	12.14
CC2011-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-04	297.83	139.16	0.47	117.18	756.97	18.22
CC2011-08	316.74	169.92	0.54	107.38	934.28	13.60
CC2011-10	518.33	299.84	0.58	160.80	1670.76	12.02
CC2012-02	116.08	114.63	0.99	15.97	843.90	6.47
CC2012-03	865.43	572.60	0.66	223.88	3345.36	9.51
CC2012-07	276.20	113.17	0.41	122.94	620.53	26.22
CC2012-10	91.93	54.44	0.59	28.87	292.77	16.72
CC2013-01	133.67	85.11	0.64	40.12	445.41	24.16
CC2013-04	66.25	68.93	1.04	9.72	451.75	9.58
CC2013-07	89.59	55.15	0.62	26.36	304.51	13.13
CC2013-11	95.67	74.75	0.78	22.01	415.84	15.24
CC2014-02	37.71	41.98	1.11	3.85	369.21	5.20
CC2014-04	20.11	21.41	1.06	2.93	137.97	10.56
CC2014-07	202.18	95.72	0.47	77.98	524.21	16.13
CC2014-11	275.41	180.48	0.66	75.77	1001.12	13.07
CC2015-01	15.69	10.11	0.64	4.49	54.89	15.48
CC2015-04	224.97	180.48	0.80	48.51	1043.37	12.12
CC2015-07	390.05	339.87	0.87	77.02	1975.39	13.12
CC2015-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-01	20.33	14.43	0.71	5.34	77.43	18.84
CC2016-04	166.42	149.91	0.90	30.51	907.76	11.00
CC2016-07	88.63	61.11	0.69	22.50	349.20	11.10
CC2016-11	25.32	18.57	0.73	6.08	105.36	12.31
CC2017-01	6.93	7.36	1.06	1.06	45.39	12.92
CC2017-04	168.85	91.47	0.54	56.09	508.33	12.32
CC2017-08	193.95	146.40	0.75	44.96	836.58	12.10
CC2017-11	143.65	148.71	1.04	21.96	939.89	11.00
CC2018-02	215.21	220.54	1.02	29.02	1595.87	7.00
CC2018-04	53.48	50.82	0.95	9.80	291.98	16.38
CC2018-06	97.17	79.64	0.82	21.41	440.99	16.98
CC2018-10	98.66	108.93	1.10	14.63	665.38	14.50
CC2019-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-07	267.02	154.17	0.58	88.69	803.94	26.18
CC2019-11	103.32	59.16	0.57	32.51	328.35	12.38
CC2020-01	161.76	90.71	0.56	53.85	485.94	17.70
CC2021-07	96.77	48.34	0.50	34.55	271.10	11.81
CC2021-11	72.44	73.10	1.01	11.20	468.50	10.02
CC2204SH	15.16	11.63	0.77	3.11	73.78	7.63
CC2208	172.90	87.94	0.51	60.81	491.63	12.01
CC2211SR	164.43	113.18	0.69	41.34	654.02	10.39
CC2301RL	2424.97	2500.40	1.03	434.53	13532.82	40.52
CC2303SH	8.66	9.45	1.09	1.27	58.96	12.73
CC2311SR	0.00	0.00	0.00	0.00	0.00	0.00
Total	163.07	49.02	0.30	91.02	292.18	104.94

Species: ER

Label	Estimate	se	cv	lcl	ucl	df
CC2004-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2004-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2006-02	2.51	1.65	0.66	0.70	8.94	15.66
CC2006-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2006-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2006-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2007-01	1.84	1.89	1.02	0.30	11.39	13.60
CC2007-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2007-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2007-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-03	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-01	2.23	2.23	1.00	0.37	13.33	13.64
CC2009-03	1.67	1.27	0.76	0.39	7.21	13.03
CC2009-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-01	1.00	1.08	1.08	0.15	6.56	14.58
CC2010-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-01	17.03	8.17	0.48	6.56	44.20	18.57
CC2011-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2012-02	2.89	2.03	0.70	0.64	13.12	6.62
CC2012-03	0.56	0.52	0.93	0.09	3.54	7.40
CC2012-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2012-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-01	3.60	2.98	0.83	0.77	16.95	13.95
CC2013-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-04	1.58	1.56	0.98	0.26	9.81	10.51
CC2014-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2015-01	13.64	10.39	0.76	3.20	58.11	14.12
CC2015-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2015-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2015-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-01	1.00	1.01	1.01	0.16	6.04	13.63
CC2016-04	0.43	0.45	1.05	0.06	2.81	11.49
CC2016-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2017-01	4.68	4.01	0.86	0.94	23.34	12.81
CC2017-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2017-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2017-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-02	2.77	2.91	1.05	0.37	20.95	7.31
CC2018-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-06	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-02	2.36	2.55	1.08	0.21	26.15	4.17
CC2019-04	1.90	1.14	0.60	0.56	6.40	11.45
CC2019-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2020-01	7.28	5.44	0.75	1.75	30.36	14.17
CC2021-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2021-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2204SH	0.00	0.00	0.00	0.00	0.00	0.00
CC2208	0.00	0.00	0.00	0.00	0.00	0.00
CC2211SR	0.00	0.00	0.00	0.00	0.00	0.00
CC2301RL	0.00	0.00	0.00	0.00	0.00	0.00
CC2303SH	1.63	1.40	0.86	0.33	8.13	12.81
CC2311SR	0.00	0.00	0.00	0.00	0.00	0.00
Total	1.01	0.28	0.27	0.59	1.72	98.66

Species: GG

Label	Estimate	se	cv	lcl	ucl	df
CC2004-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2004-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-11	7.20	7.44	1.03	1.13	45.95	12.25
CC2006-02	8.03	5.83	0.73	2.00	32.29	14.47
CC2006-04	5.59	5.76	1.03	0.91	34.51	14.23
CC2006-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2006-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2007-01	10.69	11.23	1.05	1.67	68.43	13.44
CC2007-04	34.78	35.73	1.03	5.28	229.29	10.20
CC2007-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2007-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-03	47.30	49.46	1.05	6.77	330.44	9.01
CC2008-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-03	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-07	8.58	6.21	0.72	2.16	34.03	15.64
CC2009-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-01	4.07	3.99	0.98	0.71	23.30	15.35
CC2011-04	15.87	12.34	0.78	3.64	69.10	14.42
CC2011-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-10	32.92	26.94	0.82	7.04	153.88	13.39
CC2012-02	76.95	52.03	0.68	17.50	338.39	6.36
CC2012-03	0.00	0.00	0.00	0.00	0.00	0.00
CC2012-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2012-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-01	4.05	4.01	0.99	0.68	24.02	13.31
CC2013-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-07	13.62	14.04	1.03	2.18	85.29	13.26
CC2013-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-04	5.75	5.81	1.01	0.89	37.04	10.19
CC2014-07	25.99	23.44	0.90	5.05	133.82	15.60
CC2014-11	3.81	3.65	0.96	0.67	21.56	13.64
CC2015-01	3.36	3.37	1.00	0.55	20.32	13.31
CC2015-04	135.97	135.25	0.99	22.32	828.35	12.00
CC2015-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2015-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-04	23.25	19.12	0.82	4.79	112.89	11.09
CC2016-07	68.93	65.41	0.95	11.83	401.69	11.06
CC2016-11	24.41	25.34	1.04	3.75	158.77	11.50
CC2017-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2017-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2017-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2017-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-02	9.44	7.02	0.74	2.00	44.64	7.34
CC2018-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-06	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-10	23.52	17.56	0.75	5.57	99.27	12.82
CC2019-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-07	2.59	2.72	1.05	0.40	16.89	12.28
CC2019-11	8.57	6.69	0.78	1.86	39.55	10.43
CC2020-01	30.76	30.49	0.99	5.17	183.18	13.21
CC2021-07	15.66	9.75	0.62	4.46	54.92	11.39
CC2021-11	54.91	61.12	1.11	8.08	373.31	14.73
CC2204SH	0.00	0.00	0.00	0.00	0.00	0.00
CC2208	6.15	6.24	1.01	0.99	38.32	12.21
CC2211SR	0.00	0.00	0.00	0.00	0.00	0.00
CC2301RL	6.33	6.66	1.05	0.87	46.04	8.17
CC2303SH	0.00	0.00	0.00	0.00	0.00	0.00
CC2311SR	0.00	0.00	0.00	0.00	0.00	0.00
Total	10.27	2.85	0.28	5.94	17.76	48.24

Species: LO

Label	Estimate	se	cv	lcl	ucl	df
CC2004-07	15.64	10.69	0.68	4.12	59.30	13.52
CC2004-11	38.78	42.00	1.08	5.82	258.29	13.50
CC2005-01	3.53	3.48	0.99	0.59	21.23	12.11
CC2005-04	85.56	51.56	0.60	26.02	281.33	14.45
CC2005-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-11	1.29	1.33	1.03	0.20	8.21	12.11
CC2006-02	66.78	53.01	0.79	14.93	298.74	14.11
CC2006-04	100.37	94.68	0.94	18.21	553.35	14.36
CC2006-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2006-10	1.16	1.26	1.08	0.17	7.93	12.09
CC2007-01	21.97	22.44	1.02	3.56	135.62	13.43
CC2007-04	71.81	67.72	0.94	12.29	419.75	10.51
CC2007-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2007-11	54.71	57.16	1.04	8.56	349.66	13.07
CC2008-01	109.73	106.09	0.97	19.26	625.09	14.21
CC2008-03	22.11	20.10	0.91	3.88	126.18	9.49
CC2008-08	18.02	17.54	0.97	3.09	104.92	13.13
CC2008-10	7.25	7.30	1.01	1.19	44.15	13.12
CC2009-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-03	12.61	8.27	0.66	3.48	45.62	13.69
CC2009-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-11	49.36	49.89	1.01	7.99	305.09	12.40
CC2011-01	41.90	44.00	1.05	6.69	262.27	15.33
CC2011-04	87.35	58.67	0.67	23.84	320.04	15.29
CC2011-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-10	6.66	6.87	1.03	1.05	42.30	12.39
CC2012-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2012-03	0.00	0.00	0.00	0.00	0.00	0.00
CC2012-07	112.39	101.29	0.90	21.96	575.18	16.23
CC2012-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-01	22.09	20.92	0.95	3.93	124.14	13.17
CC2013-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-07	93.06	92.24	0.99	15.65	553.35	13.32
CC2013-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-04	76.64	60.80	0.79	16.27	360.99	10.35
CC2014-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2015-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2015-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2015-07	6.96	7.15	1.03	1.12	43.22	13.55
CC2015-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-01	32.74	33.37	1.02	5.37	199.62	14.22
CC2016-04	94.90	47.32	0.50	34.09	264.21	12.30
CC2016-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-11	4.91	5.08	1.03	0.76	31.89	11.35
CC2017-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2017-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2017-08	37.19	37.30	1.00	6.15	225.05	13.16
CC2017-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-06	25.11	24.90	0.99	4.24	148.89	13.60
CC2018-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-04	168.66	120.31	0.71	41.48	685.76	11.61
CC2019-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2020-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2021-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2021-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2204SH	90.42	96.37	1.07	11.67	700.75	7.21
CC2208	0.00	0.00	0.00	0.00	0.00	0.00
CC2211SR	0.00	0.00	0.00	0.00	0.00	0.00
CC2301RL	0.00	0.00	0.00	0.00	0.00	0.00
CC2303SH	515.72	535.20	1.04	80.70	3295.87	12.48
CC2311SR	0.00	0.00	0.00	0.00	0.00	0.00
Total	29.96	9.14	0.30	16.22	55.36	25.13

Species: MN

Label	Estimate	se	cv	lcl	ucl	df
CC2004-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2004-11	2.02	1.25	0.62	0.59	6.88	13.69
CC2005-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-04	3.28	3.27	1.00	0.55	19.64	13.31
CC2005-07	0.35	0.35	1.00	0.06	2.07	13.46
CC2005-11	3.19	3.04	0.95	0.56	18.26	12.45
CC2006-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2006-04	2.68	1.81	0.68	0.73	9.87	15.43
CC2006-07	0.29	0.30	1.01	0.05	1.76	14.40
CC2006-10	0.23	0.24	1.02	0.04	1.46	12.18
CC2007-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2007-04	0.82	0.85	1.03	0.12	5.43	10.49
CC2007-07	2.33	2.49	1.07	0.36	15.11	14.33
CC2007-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-03	4.35	4.49	1.03	0.64	29.66	9.24
CC2008-08	4.80	5.11	1.07	0.73	31.37	13.29
CC2008-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-01	1.99	1.30	0.66	0.55	7.17	14.00
CC2009-03	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-01	0.67	0.65	0.98	0.12	3.83	15.23
CC2010-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-01	0.72	0.74	1.02	0.12	4.29	17.62
CC2011-04	0.28	0.28	1.03	0.05	1.70	14.40
CC2011-08	1.01	1.02	1.01	0.17	6.13	13.45
CC2011-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2012-02	0.83	0.79	0.95	0.12	5.79	6.28
CC2012-03	0.00	0.00	0.00	0.00	0.00	0.00
CC2012-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2012-10	0.75	0.78	1.03	0.12	4.71	13.31
CC2013-01	3.86	2.29	0.59	1.18	12.56	14.04
CC2013-04	11.09	8.30	0.75	2.50	49.15	9.89
CC2013-07	0.84	0.63	0.75	0.20	3.52	14.54
CC2013-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-04	6.86	4.45	0.65	1.85	25.38	10.75
CC2014-07	0.79	0.83	1.05	0.13	4.96	15.36
CC2014-11	2.15	0.98	0.45	0.86	5.41	15.08
CC2015-01	2.17	1.34	0.62	0.64	7.37	13.80
CC2015-04	2.27	1.60	0.71	0.58	8.95	13.27
CC2015-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2015-11	1.53	1.26	0.82	0.33	7.20	13.57
CC2016-01	0.29	0.30	1.04	0.05	1.84	13.23
CC2016-04	0.76	0.42	0.55	0.25	2.34	11.73
CC2016-07	0.24	0.24	0.98	0.04	1.46	11.26
CC2016-11	1.36	1.14	0.84	0.29	6.36	16.94
CC2017-01	1.07	0.82	0.77	0.24	4.70	12.46
CC2017-04	2.06	1.24	0.60	0.62	6.87	12.60
CC2017-08	0.22	0.23	1.05	0.03	1.42	12.25
CC2017-11	2.13	2.17	1.02	0.33	13.56	11.26
CC2018-02	1.64	0.98	0.60	0.45	5.96	7.39
CC2018-04	4.09	3.50	0.86	0.82	20.41	12.55
CC2018-06	1.85	1.84	1.00	0.31	11.00	13.98
CC2018-10	0.93	0.72	0.77	0.21	4.10	12.91
CC2019-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-04	1.18	0.78	0.66	0.31	4.46	10.48
CC2019-07	4.67	5.10	1.09	0.69	31.78	12.83
CC2019-11	0.84	0.84	1.00	0.13	5.33	10.35
CC2020-01	2.23	1.21	0.54	0.75	6.61	14.09
CC2021-07	2.55	2.74	1.08	0.37	17.41	11.36
CC2021-11	1.02	1.05	1.03	0.15	6.77	10.18
CC2204SH	2.72	2.65	0.97	0.40	18.38	7.39
CC2208	3.32	3.23	0.97	0.56	19.55	12.39
CC2211SR	29.05	29.21	1.01	4.54	186.02	10.23
CC2301RL	0.45	0.48	1.05	0.06	3.29	8.24
CC2303SH	36.22	27.19	0.75	8.49	154.49	12.49
CC2311SR	7.85	7.07	0.90	1.49	41.24	13.46
Total	2.44	0.64	0.26	1.45	4.11	33.97

Species: PD

Label	Estimate	se	cv	lcl	ucl	df
CC2004-07	1.57	1.52	0.97	0.27	9.10	13.24
CC2004-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-01	6.14	5.99	0.98	1.02	36.83	11.48
CC2005-04	50.88	22.43	0.44	20.60	125.65	13.95
CC2005-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-11	18.03	17.19	0.95	3.15	103.21	12.45
CC2006-02	8.93	9.17	1.03	1.45	54.96	14.21
CC2006-04	35.48	23.31	0.66	9.86	127.76	14.46
CC2006-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2006-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2007-01	30.71	21.04	0.69	8.07	116.85	13.39
CC2007-04	55.42	33.18	0.60	16.25	189.03	10.39
CC2007-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2007-11	6.14	6.23	1.01	0.99	38.14	12.40
CC2008-01	6.75	6.99	1.04	1.07	42.58	13.19
CC2008-03	68.47	38.67	0.56	21.03	222.92	9.53
CC2008-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-03	16.53	10.07	0.61	4.89	55.89	12.59
CC2009-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-01	70.73	46.81	0.66	19.65	254.62	15.52
CC2011-04	18.72	12.81	0.68	4.99	70.25	14.81
CC2011-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2012-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2012-03	111.98	70.77	0.63	28.91	433.79	7.45
CC2012-07	2.12	2.19	1.03	0.35	12.93	15.53
CC2012-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-01	30.83	22.46	0.73	7.57	125.64	13.44
CC2013-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-02	4.97	5.04	1.01	0.58	42.64	5.09
CC2014-04	15.65	8.95	0.57	4.84	50.65	10.73
CC2014-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-11	6.89	6.96	1.01	1.13	42.05	13.22
CC2015-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2015-04	24.08	23.27	0.97	4.12	140.78	12.23
CC2015-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2015-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2017-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2017-04	4.21	4.26	1.01	0.68	26.14	12.19
CC2017-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2017-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-06	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-07	2.73	2.94	1.08	0.40	18.44	12.18
CC2019-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2020-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2021-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2021-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2204SH	0.00	0.00	0.00	0.00	0.00	0.00
CC2208	0.00	0.00	0.00	0.00	0.00	0.00
CC2211SR	4.18	3.08	0.74	0.98	17.92	10.64
CC2301RL	0.00	0.00	0.00	0.00	0.00	0.00
CC2303SH	0.00	0.00	0.00	0.00	0.00	0.00
CC2311SR	0.00	0.00	0.00	0.00	0.00	0.00
Total	8.60	1.77	0.21	5.72	12.94	56.09

Species: TT

Label	Estimate	se	cv	lcl	ucl	df
CC2004-07	26.25	23.07	0.88	5.15	133.78	13.45
CC2004-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-04	1.98	1.94	0.98	0.34	11.61	13.36
CC2005-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-11	76.44	72.35	0.95	13.46	434.21	12.36
CC2006-02	7.27	7.09	0.97	1.27	41.78	14.40
CC2006-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2006-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2006-10	8.61	9.32	1.08	1.27	58.41	12.27
CC2007-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2007-04	21.41	22.02	1.03	3.25	141.22	10.25
CC2007-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2007-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-03	50.72	53.34	1.05	7.25	354.78	9.22
CC2008-08	50.07	52.08	1.04	7.91	316.98	13.32
CC2008-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-01	158.76	157.78	0.99	26.62	946.86	13.35
CC2009-03	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-11	241.91	214.24	0.89	46.38	1261.93	12.41
CC2011-01	26.47	15.82	0.60	8.21	85.36	16.16
CC2011-04	3.07	3.26	1.06	0.48	19.71	14.33
CC2011-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-10	9.70	9.99	1.03	1.53	61.58	12.30
CC2012-02	16.75	17.52	1.05	2.07	135.59	6.15
CC2012-03	0.00	0.00	0.00	0.00	0.00	0.00
CC2012-07	25.64	15.05	0.59	8.10	81.16	16.21
CC2012-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-02	141.11	164.03	1.16	13.28	1499.06	5.10
CC2014-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-07	33.26	34.59	1.04	5.38	205.49	15.37
CC2014-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2015-01	5.64	5.88	1.04	0.89	35.77	13.32
CC2015-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2015-07	35.13	26.61	0.76	8.26	149.47	13.62
CC2015-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-04	8.64	9.01	1.04	1.31	56.77	11.27
CC2016-07	137.04	104.43	0.76	31.15	602.78	11.52
CC2016-11	216.11	223.28	1.03	33.30	1402.47	11.28
CC2017-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2017-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2017-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2017-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-06	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-10	10.52	11.33	1.08	1.56	70.84	12.28
CC2019-02	242.39	261.17	1.08	21.65	2713.99	4.09
CC2019-04	48.12	53.04	1.10	6.64	348.90	10.22
CC2019-07	7.08	7.37	1.04	1.10	45.55	12.30
CC2019-11	86.13	88.13	1.02	13.15	564.20	10.26
CC2020-01	152.95	137.46	0.90	29.17	801.89	13.43
CC2021-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2021-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2204SH	0.00	0.00	0.00	0.00	0.00	0.00
CC2208	9.61	9.77	1.02	1.54	59.96	12.31
CC2211SR	3.59	2.64	0.73	0.84	15.36	10.51
CC2301RL	0.00	0.00	0.00	0.00	0.00	0.00
CC2303SH	61.58	65.87	1.07	9.23	410.96	12.28
CC2311SR	28.61	26.12	0.91	5.35	153.02	13.42
Total	27.89	8.12	0.29	15.76	49.36	57.61

Species: UC

Label	Estimate	se	cv	lcl	ucl	df
CC2004-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2004-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-01	0.74	0.54	0.73	0.19	2.91	17.80
CC2005-04	0.65	0.49	0.76	0.16	2.68	18.01
CC2005-07	5.18	3.55	0.68	1.46	18.45	27.61
CC2005-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2006-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2006-04	1.28	1.07	0.84	0.28	5.94	17.84
CC2006-07	1.07	0.83	0.77	0.26	4.48	20.17
CC2006-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2007-01	1.82	1.44	0.79	0.42	7.93	16.89
CC2007-04	0.97	0.71	0.73	0.24	3.90	14.95
CC2007-07	4.50	3.54	0.79	1.05	19.22	19.05
CC2007-11	7.40	4.54	0.61	2.29	23.90	22.17
CC2008-01	4.47	2.57	0.58	1.48	13.49	23.92
CC2008-03	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-10	1.75	1.82	1.04	0.28	10.87	15.12
CC2009-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-03	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-01	0.79	0.88	1.12	0.12	5.24	17.11
CC2010-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2012-02	4.65	4.56	0.98	0.67	32.13	7.14
CC2012-03	2.65	2.92	1.10	0.34	20.55	8.15
CC2012-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2012-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-01	3.38	3.06	0.91	0.66	17.33	16.77
CC2013-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-07	1.08	0.83	0.77	0.26	4.50	19.32
CC2014-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2015-01	0.37	0.38	1.02	0.06	2.23	15.51
CC2015-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2015-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2015-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-11	1.44	1.74	1.21	0.20	10.35	21.64
CC2017-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2017-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2017-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2017-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-06	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2020-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2021-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2021-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2204SH	0.00	0.00	0.00	0.00	0.00	0.00
CC2208	0.00	0.00	0.00	0.00	0.00	0.00
CC2211SR	0.00	0.00	0.00	0.00	0.00	0.00
CC2301RL	0.00	0.00	0.00	0.00	0.00	0.00
CC2303SH	0.00	0.00	0.00	0.00	0.00	0.00
CC2311SR	0.00	0.00	0.00	0.00	0.00	0.00
Total	0.63	0.22	0.35	0.32	1.24	94.13

Species: UD

Label	Estimate	se	cv	lcl	ucl	df
CC2004-07	32.36	28.75	0.89	6.46	162.01	16.84
CC2004-11	9.54	8.47	0.89	1.94	46.90	19.54
CC2005-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-04	2.63	2.73	1.04	0.42	16.66	13.04
CC2005-07	1.60	1.68	1.05	0.25	10.29	13.00
CC2005-11	48.51	33.50	0.69	12.55	187.53	12.72
CC2006-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2006-04	14.43	17.98	1.25	2.01	103.52	32.66
CC2006-07	18.32	14.83	0.81	4.12	81.50	17.81
CC2006-10	1.59	1.66	1.04	0.26	9.74	16.39
CC2007-01	2.64	2.67	1.01	0.43	16.22	13.01
CC2007-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2007-07	43.27	26.55	0.61	12.96	144.52	14.83
CC2007-11	1118.55	1124.48	1.01	181.17	6905.76	12.03
CC2008-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-03	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-08	5.90	4.85	0.82	1.25	27.89	13.00
CC2008-10	6.85	4.28	0.62	1.99	23.55	13.44
CC2009-01	24.02	23.02	0.96	4.20	137.31	13.04
CC2009-03	1.06	1.09	1.03	0.17	6.76	12.00
CC2009-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-01	3.29	3.30	1.00	0.56	19.47	14.87
CC2010-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-04	5.66	5.95	1.05	0.93	34.43	19.05
CC2011-08	17.18	18.14	1.06	2.67	110.37	13.72
CC2011-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2012-02	33.89	33.36	0.98	4.66	246.63	6.39
CC2012-03	6.20	7.40	1.19	0.87	44.36	19.37
CC2012-07	9.08	6.68	0.74	2.30	35.82	19.98
CC2012-10	1.43	1.45	1.02	0.24	8.58	15.52
CC2013-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2013-07	21.05	23.90	1.14	3.14	141.06	19.40
CC2013-11	1.82	1.82	1.00	0.30	11.03	13.04
CC2014-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2014-07	3.44	3.54	1.03	0.57	20.87	15.88
CC2014-11	32.02	28.62	0.89	6.12	167.36	13.12
CC2015-01	9.81	6.20	0.63	2.81	34.30	13.12
CC2015-04	52.97	38.27	0.72	13.04	215.14	12.87
CC2015-07	37.14	25.53	0.69	9.76	141.27	13.79
CC2015-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-01	1.87	2.10	1.12	0.28	12.60	16.92
CC2016-04	13.67	15.02	1.10	2.02	92.33	13.84
CC2016-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2016-11	0.49	0.52	1.06	0.07	3.33	11.00
CC2017-01	4.72	5.82	1.23	0.66	33.62	30.07
CC2017-04	16.43	13.55	0.83	3.64	74.19	19.10
CC2017-08	11.19	10.89	0.97	1.89	66.26	12.02
CC2017-11	9.51	9.76	1.03	1.49	60.74	11.63
CC2018-02	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2018-06	0.99	1.05	1.06	0.16	6.22	16.77
CC2018-10	5.82	5.86	1.01	0.95	35.65	12.71
CC2019-02	31.48	23.72	0.75	6.53	151.67	7.30
CC2019-04	69.15	71.74	1.04	11.15	428.85	14.73
CC2019-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2019-11	269.33	174.69	0.65	72.80	996.42	10.73
CC2020-01	436.83	354.00	0.81	101.31	1883.56	25.69
CC2021-07	12.94	12.22	0.94	2.23	75.01	11.00
CC2021-11	18.51	19.24	1.04	2.79	122.93	10.56
CC2204SH	0.00	0.00	0.00	0.00	0.00	0.00
CC2208	2.11	2.19	1.04	0.33	13.59	12.00
CC2211SR	9.93	6.12	0.62	2.84	34.78	10.72
CC2301RL	0.00	0.00	0.00	0.00	0.00	0.00
CC2303SH	0.00	0.00	0.00	0.00	0.00	0.00
CC2311SR	17.29	15.18	0.88	3.38	88.44	13.06
Total	35.69	17.39	0.49	13.45	94.74	16.42

Species: ULW

Label	Estimate	se	cv	lcl	ucl	df
CC2004-07	2.35	1.01	0.43	0.97	5.68	13.62
CC2004-11	1.81	0.56	0.31	0.95	3.46	14.22
CC2005-01	0.24	0.24	0.97	0.04	1.46	11.06
CC2005-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2005-07	0.96	0.59	0.61	0.29	3.23	13.23
CC2005-11	1.24	0.61	0.49	0.46	3.39	13.36
CC2006-02	1.90	0.97	0.51	0.69	5.25	15.86
CC2006-04	0.57	0.57	1.01	0.09	3.40	14.12
CC2006-07	0.00	0.00	0.00	0.00	0.00	0.00
CC2006-10	1.92	1.32	0.69	0.50	7.40	12.21
CC2007-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2007-04	0.00	0.00	0.00	0.00	0.00	0.00
CC2007-07	2.44	1.94	0.80	0.54	10.97	14.14
CC2007-11	1.35	1.29	0.96	0.23	7.84	12.06
CC2008-01	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-03	0.00	0.00	0.00	0.00	0.00	0.00
CC2008-08	0.82	0.50	0.61	0.25	2.75	13.53
CC2008-10	0.00	0.00	0.00	0.00	0.00	0.00
CC2009-01	0.85	0.68	0.80	0.19	3.87	13.00
CC2009-03	0.56	0.36	0.64	0.16	2.02	12.07
CC2009-07	0.44	0.42	0.97	0.08	2.47	15.23
CC2009-11	0.28	0.30	1.05	0.04	1.87	11.10
CC2010-01	7.52	3.83	0.51	2.73	20.70	16.90
CC2010-08	0.00	0.00	0.00	0.00	0.00	0.00
CC2010-11	0.00	0.00	0.00	0.00	0.00	0.00
CC2011-01	0.71	0.74	1.05	0.11	4.42	15.30
CC2011-04	6.01	2.76	0.46	2.38	15.16	16.68
CC2011-08	2.54	1.53	0.60	0.77	8.39	13.57
CC2011-10	1.70	0.98	0.58	0.53	5.42	12.31
CC2012-02	4.17	2.75	0.66	0.96	18.04	6.08
CC2012-03	5.97	3.59	0.60	1.61	22.07	7.15
CC2012-07	2.07	1.07	0.52	0.74	5.83	15.38
CC2012-10	5.04	2.92	0.58	1.58	16.07	13.26
CC2013-01	0.77	0.80	1.03	0.12	4.85	13.07
CC2013-04	4.22	1.82	0.43	1.66	10.70	9.36
CC2013-07	6.55	2.47	0.38	3.00	14.30	14.54
CC2013-11	3.97	1.61	0.41	1.71	9.22	13.34
CC2014-02	1.24	0.82	0.66	0.27	5.80	5.08
CC2014-04	4.12	1.60	0.39	1.79	9.48	10.32
CC2014-07	3.85	2.27	0.59	1.21	12.30	15.28
CC2014-11	1.69	1.14	0.68	0.45	6.33	13.13
CC2015-01	1.87	1.19	0.64	0.53	6.58	13.08
CC2015-04	1.98	1.52	0.77	0.45	8.67	12.26
CC2015-07	3.96	1.76	0.44	1.59	9.86	13.36
CC2015-11	0.84	0.51	0.61	0.25	2.81	13.31
CC2016-01	3.97	2.45	0.62	1.17	13.51	13.29
CC2016-04	3.59	2.19	0.61	1.04	12.32	11.21
CC2016-07	4.08	3.62	0.89	0.76	21.77	11.14
CC2016-11	1.83	1.22	0.67	0.48	6.96	11.13
CC2017-01	1.75	1.19	0.68	0.46	6.70	12.15
CC2017-04	1.47	1.06	0.72	0.36	5.96	12.41
CC2017-08	3.17	1.36	0.43	1.30	7.71	12.46
CC2017-11	1.54	1.12	0.72	0.37	6.42	11.12
CC2018-02	3.23	2.24	0.69	0.74	14.15	7.10
CC2018-04	0.52	0.40	0.77	0.12	2.32	12.16
CC2018-06	4.46	1.86	0.42	1.89	10.54	13.64
CC2018-10	7.09	3.73	0.53	2.44	20.65	12.78
CC2019-02	1.22	0.91	0.75	0.19	7.69	4.05
CC2019-04	1.40	0.97	0.69	0.35	5.63	10.10
CC2019-07	2.43	0.87	0.36	1.15	5.12	13.76
CC2019-11	0.89	0.89	1.00	0.14	5.67	10.12
CC2020-01	3.69	1.62	0.44	1.50	9.09	13.57
CC2021-07	5.81	2.34	0.40	2.48	13.61	11.27
CC2021-11	4.04	2.36	0.58	1.21	13.45	10.20
CC2204SH	12.52	12.52	1.00	1.76	89.11	7.10
CC2208	1.60	1.06	0.66	0.43	5.95	12.01
CC2211SR	6.52	4.52	0.69	1.63	26.17	10.29
CC2301RL	3.18	3.01	0.95	0.52	19.48	8.89
CC2303SH	1.54	0.96	0.62	0.45	5.33	12.17
CC2311SR	1.96	1.62	0.82	0.42	9.27	13.18
Total	2.46	0.30	0.12	1.92	3.14	56.13

7 References

• Campbell et al. (2015) Inter-annual and seasonal trends in cetacean distribution, density and abundance off southern California. doi:10.1016/j.dsr2.2014.10.008
• Lerzack and Hobbs (1997) Calculating sighting distances from angular readings during shipboard, aerial, and shore-based marine mamaml surveys. doi:10.1111/j.1748-7692.1998.tb00745.x