Do Non-masting Red Maple Species Exhibt Muted Dynamics in Comparison to Masting Sugar Maple Species
Sydney Shenk
Introduction
This report seeks to answer the question: Does the non-masting Red Maple exhibit muted dynamics in comparison to the masting Sugar Maple. Masting tree species have periodic years where they will produce an overabundance of seeds/fruits. Contrastly, non-masting species will produce a similar number of seeds/fruits ever year with not much change. Within botany, dynamics refers the the processes and interactions of the plants that contribute their their overall growth and survival within their ecosystem. Tree dynamics can be monitored through factors such as flower production, seed production, leaf production, sap production, size, etc. which all give insights into whether a tree is healthy or having a more successful reproduction year. Muted dynamics then means that these variables of tree health are depressed or lesser than other trees with more pronounced dynamics.
Libraries
Loaded below are the libraries necessary to complete this analysis.
library(tidyverse)
library(modelr)
library(readr)tidyverse: includes basic functions and functions to make visualizations
modelr: includes functions to create models based on data
readr: includes functions to import data from csv files
Data
The data used in this analysis was taken from a collection of data
done at Harvard Forest monitoring maple reproduction and sap flow. The
full data set can be found at https://portal.edirepository.org/nis/mapbrowse?scope=knb-lter-hfr&identifier=285&revision=6.
There are 14 total data sets available within this collection, but based
on the parameters of the question I have chosen to include only three
within my analysis. The maple_seed data set will be used to
determine which years are masting years and which are nonmasting years
for the sugar maple trees. The maple_tap and
maple_sapwill be used to compare the dynamics between the
Red Maple and the Sugar Maple looking at the circumference of the trunk,
amount of sap produced, and the concentration of sugar in the sap.
maple_tap <- read_csv("hf285-01-maple-tap.csv")
maple_sap <- read_csv("hf285-02-maple-sap.csv")
maple_seed <- read_csv("hf285-09-maple-seed-count.csv")
datatable(maple_tap)datatable(maple_sap)datatable(maple_seed)Unfortunately it is not possible to combine data sets without
inducing a significant number of NAs or severely limited the data
available. Because of this, data from maple_tap,
maple_sap, and maple_seed will be shown
separately.
Citation upon authors request:
Rapp, J., E. Crone, and K. Stinson. 2023. Maple Reproduction and Sap Flow at Harvard Forest since 2011 ver 6. Environmental Data Initiative. https://doi.org/10.6073/pasta/7c2ddd7b75680980d84478011c5fbba9 (Accessed 2024-12-09).
maple_seed Variables (Sugar Maples Only)
date: date data was collected
tree: tree identification number
EC.count: Elizabeth Crone’s samara pairs count
JC.count: Josh Rapp’s samara pairs count
total.count: combined EC.count and JC.count, gives total count of samara pairs
note: notes for data collection
maple_sap Variables
date: date data was collected
tree: tree identification number (HR for Sugar Maple, HFR for Red Maple)
tap: designates which tap data was collected from in the same tree (A or B)
time: time of sap collection
datetime: time and date of sap collection
sugar: sugar concentration of sap collected measured in Brixx (weight percent)
species: tree species (ACSU for Sugar Maple, ACRU for Red Maple)
sap.wt: weight of sap collected (kg)
maple_tap Variables
date: date data was collected
tree: tree identification number (HR for Sugar Maple, AR for Red Maple)
tap: designates which tap data was collected from in the same tree (A or B)
species: tree species (ACSU for Sugar Maple, ACRU for Red Maple)
dbh: diameter of trunk 1.4 meters above the ground (centimeter)
tap.bearing: compass bearing of tap hole (degrees)
tap.height: height at which tap was placed above the ground (centimeter)
Data Clean Up
maple_seed Clean Up
To clean up maple_seed, date was separated into year,
month, and day columns to make analysis easier going forwards. Columns
that will not be useful for analysis such as note, EC.count, and
JR.count were removed as a matter of conciseness.
##clean up of maple_seed
summary(maple_seed)## tree date EC.count JR.count
## Length:220 Min. :2011-08-04 Min. : 0.000 Min. : 0.000
## Class :character 1st Qu.:2013-09-11 1st Qu.: 0.000 1st Qu.: 0.000
## Mode :character Median :2016-10-03 Median : 0.000 Median : 0.000
## Mean :2016-09-10 Mean : 8.068 Mean : 8.673
## 3rd Qu.:2019-09-07 3rd Qu.:16.500 3rd Qu.:18.000
## Max. :2021-09-15 Max. :39.000 Max. :40.000
## total.count note
## Min. : 0.00 Length:220
## 1st Qu.: 0.00 Class :character
## Median : 0.00 Mode :character
## Mean :16.74
## 3rd Qu.:35.25
## Max. :77.00seed_clean <- maple_seed %>%
separate(date, into = c("year", "month", "day"), sep = "-") %>%
select(-note, -EC.count, -JR.count)
summary(seed_clean)## tree year month day
## Length:220 Length:220 Length:220 Length:220
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
## total.count
## Min. : 0.00
## 1st Qu.: 0.00
## Median : 0.00
## Mean :16.74
## 3rd Qu.:35.25
## Max. :77.00datatable(seed_clean)maple_sap Clean Up
To clean up maple_sap, once again date was separated
into year, month, and day,columns to make analysis easier going
forwards. The species names were change from ACSA and ACRU to Sugar
Maple and Red Maple respectively. NAs were removed from the sugar, and
sap.wt columns. A likely data entry error in the sugar column (sugar =
22.00) was detected and removed. Finally, the time and date time columns
were removed since they will not be used throughout this analysis.
#clean up of maple_sap
sap_clean <- maple_sap %>%
#separate date into three separate columns
separate(date, into = c("year", "month", "day"), sep = "-") %>%
#change acronyms for tree species into common names
mutate(species = str_replace(species, "ACSA" , "Sugar Maple")) %>%
mutate(species = str_replace(species, "ACRU", "Red Maple")) %>%
#filter NAs
filter(!is.na(sugar)) %>%
filter(!is.na(sap.wt)) %>%
#filter outliers
filter(sugar != 22.00) %>%
#remove unnecessary columns
select(-time, -datetime)
summary(sap_clean)## year month day tree
## Length:7575 Length:7575 Length:7575 Length:7575
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
## tap sugar species sap.wt
## Length:7575 Min. :0.800 Length:7575 Min. : 0.010
## Class :character 1st Qu.:2.000 Class :character 1st Qu.: 1.840
## Mode :character Median :2.400 Mode :character Median : 3.500
## Mean :2.469 Mean : 4.224
## 3rd Qu.:2.900 3rd Qu.: 5.900
## Max. :7.300 Max. :24.040datatable(sap_clean)maple_tap Clean Up
To clean up `maple_tap`, once again date was separated into year, month, and day, columns to make analysis easier going forwards and the species acronyms were changed from ACSA and ACRU to Sugar Maple and Red Maple. The tree names were also changed from “HFR” to “AR” so the names were consistent between maple_tap and maple_sap. NAs were cleared from the data set by filtering out all NAs from the dbh column. Lastly, the tap.bearing, and tap.height columns were removed as they will not be used in the analysis going forwards.
#clean up of maple_tap, unfortunatley data for trunk width is collected only in 2012, 2016, and 2017
tap_clean <- maple_tap %>%
#separate date into three separate columns
separate(date, into = c("year", "month", "day"), sep = "-") %>%
#change acronyms for tree species into common names
mutate(species = str_replace(species, "ACSA" , "Sugar Maple")) %>%
mutate(species = str_replace(species, "ACRU", "Red Maple")) %>%
#change tree names so they are consistent between data sets
mutate(tree = str_replace(tree, "HFR1", "AR1")) %>%
mutate(tree = str_replace(tree, "HFR2", "AR2")) %>%
mutate(tree = str_replace(tree, "HFR3", "AR3")) %>%
mutate(tree = str_replace(tree, "HFR4", "AR4")) %>%
mutate(tree = str_replace(tree, "HFR5", "AR5")) %>%
mutate(tree = str_replace(tree, "HFR6", "AR6")) %>%
mutate(tree = str_replace(tree, "HFR7", "AR7")) %>%
mutate(tree = str_replace(tree, "HFR8", "AR8")) %>%
mutate(tree = str_replace(tree, "HFR9", "AR9")) %>%
#remove NAs
filter(!is.na(dbh)) %>%
select(-tap.bearing, -tap.height)
summary(tap_clean)## year month day tree
## Length:126 Length:126 Length:126 Length:126
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
## tap species dbh
## Length:126 Length:126 Min. :31.00
## Class :character Class :character 1st Qu.:54.70
## Mode :character Mode :character Median :63.85
## Mean :62.87
## 3rd Qu.:74.88
## Max. :86.20datatable(tap_clean)Analysis
Preliminary Analysis
To gain insights into variables that would be worth further investigation I created a new data set which averages a trees sap weight, sugar concentration, trunk width, and gives the minimum and maximum trunk width for each tree across the course of the data collected . It also considers how much a trees circumference has grown between the years 2016-2017 (These two years were chosen because data for dbh was only collected for both tree species in these two years). The year 2012 has been filtered out because it contains only data for Sugar Maples.
This data set contains several new variables…
mean_sap : mean sap weight for each tree (kg)
mean_sugar: mean sugar concentration for each tree (Brixx)
max_trunk_width: max circumference of trees trunk (cm)
min_trunk_width: min circumference of trees trunk (cm)
mean_trunk_width: mean circumference of trees trunk (cm)
oneyr_growth: growth of trunk circumference between the years 2016-2017 (cm)
#This data set looks at the average sap weight, sugar concentration, and trunk width
averages_tap_sap_data <- sap_clean %>%
full_join(tap_clean, by = c("tree", "year", "month", "day", "species")) %>%
group_by(tree) %>%
filter(year != 2012) %>%
summarize(mean_sap = mean(sap.wt, na.rm = TRUE),
mean_sugar = mean(sugar, na.rm = TRUE),
max_trunk_width = max(dbh, na.rm = TRUE),
min_trunk_width = min(dbh, na.rm=TRUE),
mean_trunk_width = mean(dbh, na.rm = TRUE),
oneyr_growth = max_trunk_width - min_trunk_width) %>%
separate(tree, into = c("species", "number"), sep = 2) %>%
mutate(species = case_when(species == "AR" ~ "Red Maple",
species == "HF" ~ "Sugar Maple"))
datatable(averages_tap_sap_data)Based on preliminary observations of available data, it appears that Red Maples do exhibit muted dynamics in comparison to Sugar Maples. The top 10 values within mean_sap, mean_sugar, mean_trunk_width, and oneyr_growth are dominated by Sugar Maple trees. The bottom 10 values in the mean_sap, mean_sugar, and mean_trunk are largely dominated by the Red Maples. Interestingly though, while the top 10 values of oneyr_growth are dominated by Sugar Maples, the tree with the largest oneyr_growth is a Red_Maple and 7 of the bottom 10 oneyr_growth values are Sugar Maple trees.
Determination of Masting Years
Because masting years change the tree dynamics, particularly in regards to reproduction, it will be useful to determine which years are masting years and which are nonmasting years. Masting years are determined by years with unusually high levels of seed production. To determine which years were masting years the total seed count produced by Sugar Maples (the masting species) was compared across years using the `maple_seed` data set.
ggplot(seed_clean) +
geom_boxplot(aes(x = year, y=total.count))
Based on this plot, it appears that 2011, 2013, 2017, & 2019 were masting years for the Sugar Maple trees. 2021 appears to have had some trees which exhibited masting years, but largely was a nonmasting year for most of the trees and as such will be included within the nonmasting category. These observations were then added into the cleaned ‘maple_sap’ and ‘maple_tap’ data set creating a new variable called ‘masting’ which has two possible variables, “masting” indicates the year was a masting year and “nonmasting” indicates the year was a nonmasting year.
#add masting_yr variable to sap_clean
sap_clean_w_masting <- sap_clean %>%
mutate(masting = case_when(year == 2011 ~ "masting",
year == 2013 ~ "masting",
year == 2017 ~ "masting",
year == 2019 ~ "masting",
year == 2012 ~ "nonmasting",
year == 2014 ~ "nonmasting",
year == 2015 ~ "nonmasting",
year == 2016 ~ "nonmasting",
year == 2018 ~ "nonmasting",
year == 2020 ~ "nonmasting",
year == 2021 ~ "nonmasting",
))%>%
filter(!is.na(masting))
datatable(sap_clean_w_masting)#add masting variable into tap_clean
tap_clean_w_masting <- tap_clean %>%
mutate(masting = case_when(year == 2011 ~ "masting",
year == 2013 ~ "masting",
year == 2017 ~ "masting",
year == 2019 ~ "masting",
year == 2012 ~ "nonmasting",
year == 2014 ~ "nonmasting",
year == 2015 ~ "nonmasting",
year == 2016 ~ "nonmasting",
year == 2018 ~ "nonmasting",
year == 2020 ~ "nonmasting",
year == 2021 ~ "nonmasting",
))%>%
filter(!is.na(masting))
datatable(tap_clean_w_masting)Comparison of sap production
Sap is used for transporting nutrients throughout a tree. As such we can better understand the dynamics and health fo trees by studying the sap they produce. The box plot below compares the amount of sap collected from taps to the tree species and the masting years.
ggplot(sap_clean_w_masting) +
geom_boxplot(aes(x = species, y = sap.wt, color = masting)) +
labs(x = "Species",
y = "Sap Weight (kg)",
color = "Masting Year")
This plot suggests that the Sugar Maple produces more sap than the Red Maple. In nonmasting years it appears that the Red Maple may produce slightly more sap and the Sugar Maple produces slightly less, but these differences are very slight and may be negligible.
To further explore the significance of these trends we can create a linear model comparing the sap weight collected to the species of the tree.
sap_lm <- lm(sap.wt ~ species, data = sap_clean_w_masting)
summary(sap_lm)##
## Call:
## lm(formula = sap.wt ~ species, data = sap_clean_w_masting)
##
## Residuals:
## Min 1Q Median 3Q Max
## -4.3497 -2.2597 -0.5997 1.5803 19.6803
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.3095 0.1100 21.00 <2e-16 ***
## speciesSugar Maple 2.0502 0.1162 17.65 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 2.997 on 7143 degrees of freedom
## Multiple R-squared: 0.04178, Adjusted R-squared: 0.04164
## F-statistic: 311.4 on 1 and 7143 DF, p-value: < 2.2e-16The p-value of a model indicates the probability that the difference between two variables is due to chance. Significant differences are determined by a p <= 0.05.
This model is significantly different from the mean model (p < 2.2e-16), however its R2 is only 0.0416 indicating that it may not be very reliable. Despite its reliability it suggests that species does have a significant impact on sap weight (p < 2e-16). On average Sugar Maple species secreted 2.050 kg sap more than Red Maple trees, and suggests that Red Maples have muted sap dynamics compared to Sugar Maples.
We can add a secondary causitive variable into the model to investigate the effect of both masting year and species on sap weight.
sap_lm_2 <- lm(sap.wt ~ masting + species, data = sap_clean_w_masting)
summary(sap_lm_2)##
## Call:
## lm(formula = sap.wt ~ masting + species, data = sap_clean_w_masting)
##
## Residuals:
## Min 1Q Median 3Q Max
## -4.3862 -2.2662 -0.5929 1.5738 19.7071
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.36357 0.12286 19.239 <2e-16 ***
## mastingnonmasting -0.07328 0.07427 -0.987 0.324
## speciesSugar Maple 2.04260 0.11643 17.544 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 2.997 on 7142 degrees of freedom
## Multiple R-squared: 0.04191, Adjusted R-squared: 0.04164
## F-statistic: 156.2 on 2 and 7142 DF, p-value: < 2.2e-16When multiple variables are present the model compares back the the intercept when all other variables are held constant.
This model once again is significantly different from the mean model (p < 2.2e-16) but it does not improve upon the RSE or R2 values of the previous model. The reliabilty of this model leaves great room for improvement with R2 = 0.0416. For sap weight collected, the masting year does not have a significant impact (p = 0.324). However, the model further confirms that tree species significantly impacts the sap excreted (p < 2e-16). It suggests that on average Sugar Maples secrete 2.043 kg sap more than Red Maples further confirming the muted dynamics of Red Maples.
Comparison of sugar production
The primary fuel of maple trees is the sugar they produce through photosynthesis. As such the concentration of sugar in their sap can be key to understanding the health of the tree and their overall dynamics. The box plot below shows the difference in sugar concentration in the sap collected from Sugar Maples and Red Maples in masting and nonmasting years.
ggplot(sap_clean_w_masting) +
geom_boxplot(aes(x = species, y = sugar, color = masting)) +
labs (x = "Sugar Concentration",
y = "Species",
color = "Masting Year")
Based on this plot it appears that Sugar Maple’s produce a higher concentration of Sugar in their sap than Red Maples in both masting and nonmasting years. Similar to the trend seen in the sap weight, for Sugar Maples in nonmasting years it seems that there may be a slight decrease in sugar concentration in than in masting years. On the opposite side, even though Red Maples do not experience masting it seems that during nonmasting years Red Maples experience a slight increase in sugar concentration. That being said, the differences within both species are slight and may not be significant.
To investigate the significance of this trend we can create a linear model. This model compares the effect of species on sugar concentration.
sugar_lm <- lm(sugar ~ species, data = sap_clean_w_masting)
summary(sugar_lm)##
## Call:
## lm(formula = sugar ~ species, data = sap_clean_w_masting)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.7539 -0.3539 -0.0539 0.3461 4.7461
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1.84293 0.02136 86.29 <2e-16 ***
## speciesSugar Maple 0.71099 0.02256 31.51 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.5821 on 7143 degrees of freedom
## Multiple R-squared: 0.1221, Adjusted R-squared: 0.1219
## F-statistic: 993.1 on 1 and 7143 DF, p-value: < 2.2e-16Based on this model, the species of tree has a significant impact on the sugar concentration in tree sap (p < 2e-16).. Sugar Maples on average have 0.711 Brixx higher than the sugar concentration in Red Maples indicating that Red Maples have muted dynamics when it comes to Sugar concentration.
The model as a whole is significantly different than the mean model with p < 2.2e-16. However, the model has a low R2 =0.1219 and is not very reliable.
To improve upon the model we can add in a secondary causitive variable to understand the impact of both masting year and species on sugar concentration.
sugar_lm_2 <- lm(sugar ~ masting + species, data = sap_clean_w_masting)
summary(sugar_lm_2)##
## Call:
## lm(formula = sugar ~ masting + species, data = sap_clean_w_masting)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.7390 -0.3798 -0.0798 0.3610 4.7202
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1.87299 0.02385 78.539 < 2e-16 ***
## mastingnonmasting -0.04075 0.01442 -2.827 0.00471 **
## speciesSugar Maple 0.70679 0.02260 31.274 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.5819 on 7142 degrees of freedom
## Multiple R-squared: 0.123, Adjusted R-squared: 0.1228
## F-statistic: 501 on 2 and 7142 DF, p-value: < 2.2e-16This model slightly improves upon the reliability of the previous model by decreasing the RSE from RSE = 0.5821 in the previous model to RSE = 0.5819 in this model. Similarly the R2 indicates a slightly more reliable increasing to R2 = 0.1228 int this model from R2 = 0.1219 in the previous model. However, these differences are rather small and the reliablity of both models still leaves great room for improvement. That being said, once again, this model is significantly different from the mean model (p < 2.2e-16).
This model further confirms that the species of the tree has a significant impact on sugar concentration (p < 2e-16). On average Sugar Maples have 0.0707 Brixx Sugar more than Red Maples, providing further evidence for the muted dynamics of Red Maple trees. The model also suggest that masting year has a significant impact on sugar concentration (p = 0.005). During nonmasting years maple trees produce an average of 0.041 Brixx Sugar less than in masting years.
Comparison of trunk width
The size of a tree can indicate their health. Typically, if the same age, healthier trees with better access to necessary nutrients will grow to be larger than their counter parts with less access. The box plot below compares the circumference of the tree trunk to the species and the masting year.
ggplot(tap_clean_w_masting) +
geom_boxplot(aes(x = species, y = dbh, color = masting)) +
labs(x = "Species",
y = "Trunk Circumference (cm)",
color = "Masting Year")
This plot suggests that Sugar Maples have a larger trunk circumference than Red Maples. It appears that in nonmasting years both Sugar Maples and Red Maples had slightly smaller trunk circumferences.
To better understand the significance of these trends we can create a linear model. The model below compares the circumference of the trunk to the species of the tree.
width_lm <- lm(dbh ~ species, data = tap_clean_w_masting)
summary(width_lm)##
## Call:
## lm(formula = dbh ~ species, data = tap_clean_w_masting)
##
## Residuals:
## Min 1Q Median 3Q Max
## -29.403 -8.648 -2.103 10.697 21.017
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 46.583 2.405 19.373 < 2e-16 ***
## speciesSugar Maple 20.120 2.673 7.528 9.19e-12 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 11.78 on 124 degrees of freedom
## Multiple R-squared: 0.3137, Adjusted R-squared: 0.3081
## F-statistic: 56.67 on 1 and 124 DF, p-value: 9.191e-12This model is significantly different from the mean model (p = 2.14e-11), but has a relatively small R2 = 0.3136 and leaves room for improvement with its reliability.
This model suggests that species has a significant impact on trunk circumference (p < 2e-16). On average Sugar Maples are 20.17 cm larger than Red Maples, providing evidence for muted dynamics in trunk circumference within Red Maple trees.
To further investigate the trends seen in the box plot we can add in a secondary causitive variable to investigate the impacts of both species and masting year on trunk circumference.
width_lm_2 <- lm(dbh ~ masting + species, data = tap_clean_w_masting)
summary(width_lm_2)##
## Call:
## lm(formula = dbh ~ masting + species, data = tap_clean_w_masting)
##
## Residuals:
## Min 1Q Median 3Q Max
## -29.013 -8.632 -2.063 10.281 21.170
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 47.136 2.647 17.805 < 2e-16 ***
## mastingnonmasting -1.106 2.183 -0.507 0.613
## speciesSugar Maple 20.282 2.700 7.513 1.03e-11 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 11.82 on 123 degrees of freedom
## Multiple R-squared: 0.3151, Adjusted R-squared: 0.304
## F-statistic: 28.3 on 2 and 123 DF, p-value: 7.772e-11Once again this model is significantly different from the mean model p = 1.81 e-10 however this p value is increased the previous model. In this model RSE is slightly increased and R2 is slightly decreased compared to the previous model indicating that it is less reliable that the previous. In this model RSE = 11.95 and R2 = 0.309. In the previous model comparing only the two variables RSE = 11.91 and R2 = 0.313.
Despite its issues with reliability, this model further confirms that species significantly impacts the circumference of the tree trunk (p < 2e-16). In this model Sugar Maples on average are 20.300 cm larger than Red Maples.
Masting year does not appear to have a significant impact on circumference (p = 0.632). This result is unsurprising due to the fact that tree trunks do not typically shrink, rather they continue to get larger year by year, varying each year in how much they grow. Becuase of this a more useful comparison may be to look at the trends in the growth of the trunk.
Comparison of growth between 2016-2017
The growth of a tree has important insights into its dynamics. Trees that have grown more than others are typically healthier than their counterparts and have more energy to put towards growth. Unfortunately, due to limitations of the data available, we can only compare the growth of trees between the years 2016 and 2017. To do so the variable oneyr_growth was created in the ‘averages_tap_sap_data’ data set by subtracting the smallest tree circumference from the largest tree circumference.
The box plot below compares species to growth between the years 2016-2017.
#compares species and growth of trunk
ggplot(averages_tap_sap_data) +
geom_boxplot(aes(x = species, y= oneyr_growth)) +
labs(x = "Species",
y = "Trunk Circumference Between 2016-2017 (cm)")
Based on this plot it seems that Sugar Maples may have grown more between 2016-2017.
To understand the significance fo this trend we can create a linear model. The model below compares species the the trunk growth between 2016-2017.
growth_model <- lm(oneyr_growth ~ species, data = averages_tap_sap_data)
summary(growth_model)##
## Call:
## lm(formula = oneyr_growth ~ species, data = averages_tap_sap_data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.8500 -0.5500 -0.4158 0.0842 5.3500
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.9500 0.3791 2.506 0.0186 *
## speciesSugar Maple -0.2342 0.4684 -0.500 0.6211
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.199 on 27 degrees of freedom
## Multiple R-squared: 0.009175, Adjusted R-squared: -0.02752
## F-statistic: 0.25 on 1 and 27 DF, p-value: 0.6211This model is not significantly different from the mean model (p = 0.6211) , indicating it is not a reliable model. Unsurprisingly, the R2 also confirms the lack of reliability with R2 = -0.0275.
If ignoring the issues the reliability of the fit of the model, the model suggests that species does not have a significant impact on trunk growth in the years data was available. However, if data is collected for more years the reliability of the model may be improved and more definite conclusions could be drawn.
Conclusions
Basd on the data available, it appears that Red Maples have muted dynamics in comparison to Sugar Maples. Sugar concentration, sap weight collected, and circumference of trunks were all significantly different between species, suggesting that the dynamics of Red Maples are more muted than Sugar Maples. Despite this trend, the R2 values for all models created using this data were relatively low and leave great room for improvement in their reliability. To address such an issue more data should be collected.
Similary, the dynamics capable of being explored were also severely limited by data availability. There was no data collected for leaves, seeds, or flowers for Red Maple trees, all of which are important variables to understand the dynamics of the species. Furthermore, data for trunk circumference for both species was only available between 2016-2017 limiting the analysis capable for tree growth. Further collection of a greater variety of data will allow better conclusions to be drawn for the difference between Sugar Maple and Red Maple dynamics.