Data Analysis for Report on Dissolved Oxygen in Great South Bay, New York, 2016-2017

1 R packages used

2 Read in tables

Raw data from NWIS was read in to RStudio using the ‘dataRetrieval’ package created by the USGS. Data was then cleaned and structured using base R and dplyr, and saved as csv in the directory “X:/transfer/RobW/TNC_DO”.

# get site.info table 
site.info <- read.csv("site_info.csv")

#get DO time series table
DO.data <- read.csv("DO_data.csv")

#merge tables
DO.data <- DO.data %>% 
  left_join(site.info,by="SiteNum")

---

3 Data Analysis and Figures

The analysis below is done in support of the document ‘TNT-manuscript template.docx’

3.1 Description of Study Area

There are six sensors located within the boundaries of two HUC12 watersheds - Browns River and Greens Creek. I have a contours layer and inset map created but gets covered up - needs work

Figure 1 - Locations of continuous DO sensors in Great South Bay. Map processed in ArcGIS v. 10.7.1.

3.2 Tables of DO averages

These tables can be included in the Appendix maybe or cited within the main text.

We see variation by year (2017 is greater than 2016), watershed (Greens is greater than Browns), by time of day (day greater than night). The table grouped by site number is depicted as a map (Figure 2)

Figure 2b can be the same map but for 2017

3.3 AUV Survey

AUV survey of GSB performed August 23-26 2016. Products (Sample points and interpolated DO surface) released on ScienceBase at https://www.sciencebase.gov/catalog/item/59284998e4b0c6c925587b35.

The interpolated surface below shows the the spatial distribution of DO in GSB under the conditions associated with 8/23/16 between 1200 and 0500.

The interpolated surface below is a heat map from the continuous sensors. Even with a dense sampling setup, interpolation misses key information. This is part fo the challenge of monitoring DO.

The map below shows areas where AUV and continuous sensor interpolations vary. Created via ArcMap ‘Raster Calculator’ tool.

3.4 Acute violations

NYSDEC Crietria: The DO concentration shall not fall below the acute standard of 3.0 mg/L at any time.

Acute fraction is defined as the number of observations below 3.0 divided by the total number of observations.

The tables show trends in DO acute violations - more in 206 than 2017, over twice as many as night, almost 3 times as many in Browns

Figure 4 - Spatial distribution of acute violations

3.5 DO mean by year/watershed/time of day

3.6 Day/Night exceedences

3.7 Day/Night exceedences by site number

Figure 2a - Mean DO for 2016

3.8 Distribution of DO values

## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

3.9 Exceedences

3.10 Water temperature

3.11 Multiple Linear Regression

3.11.1 Water temperature

## 
## Call:
## lm(formula = DOWT$minDO ~ DOWT$temprange + DOWT$TOD)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -5.1617 -1.6301 -0.0385  1.6699  4.2051 
## 
## Coefficients:
##                   Estimate Std. Error t value Pr(>|t|)    
## (Intercept)         2.8301     0.1653  17.126  < 2e-16 ***
## DOWT$temprangelow   2.3316     0.2580   9.039  < 2e-16 ***
## DOWT$temprangemid   0.4647     0.1959   2.372    0.018 *  
## DOWT$TODnight      -1.2564     0.1773  -7.087 4.45e-12 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.025 on 522 degrees of freedom
## Multiple R-squared:  0.1956, Adjusted R-squared:  0.191 
## F-statistic: 42.32 on 3 and 522 DF,  p-value: < 2.2e-16

3.12 Time Series

3.12.1 Plot Daily Minimum Values (Static and Interactive: for 2017)

DO.data %>% 
  filter(year==2016) %>%
  #left_join(site.info, by="SiteNum") %>%
  group_by(monthday, SiteNum, watershed) %>% 
  summarize(mean=mean(DO_mgL)) %>% 
  ggplot(aes(x=monthday,y=mean, color=SiteNum)) + 
    geom_line(aes(group=SiteNum), alpha=.7) + 
    stat_summary(geom='point',fun.y='mean', color="darkgreen") + 
    facet_grid(watershed~.) +
    geom_hline(yintercept = 3.0, size = 1, color = "red", linetype = "dashed")