AGU 2022: B35C-1416 Haber et al. Supporting Figures

This document contains supporting information for Haber et al.’s poster presentation at the AGU Fall Meeting 2022

Title:

Ecosystem functional recovery following disturbance: Drivers of carbon flux in a restored tidal freshwater wetland.

Abstract: Tidal freshwater wetlands (TFWs) contribute substantial uncertainty to estimates of wetland carbon (C) cycling, even as they are among the largest methane sources globally. Among the factors limiting mechanistic understanding of TFW C cycling is their susceptibility to disturbance and underrepresentation in flux data sets. In particular, we have few observations of both CH4 and CO2 from TFWs undergoing recovery following disturbance. We leverage a recently added AmeriFlux eddy covariance tower site (US-RRC) to characterize and interpret hourly to seasonal ecosystem-scale CH4 and CO2 fluxes in a recovering TFW after deforestation and decades of impoundment in the upper James River estuary, Virginia, USA. Our measurements include CH4 and CO2 fluxes along with a suite of biometeorological variables, water table height, and plant phenology to assess the principal biophysical drivers of C fluxes in this globally underrepresented ecosystem type. Our initial findings indicate that hourly to seasonal CH4 emissions correspond with daytime irradiance and secondarily water table height, while seasonal CO2 fluxes are primarily driven by water table height and soil temperature. Overall, our TFW may contribute net positive radiative forcing due to high methane emissions, consistent with observations from other recently disturbed TFWs. Continued measurements at our site and recent addition to a bicoastal network of eddy covariance towers will enable further exploration of C cycling drivers, including pulse salinity and nitrate loading, in these critically important yet understudied systems.

Session & Poster ID: B35C-1416

Part 1: Meteorological time series, June 6 - November 1 2022

List of meteorological variables currently collected at our site and associated sensors:

1. PAR (PPFD)
   • LI-190R quantum sensor
2. Global radiation
   • LI-200R pyranometer
3. Net radiation
   • Kipp & Zonen NRLite2 net radiometer
4. Rainfall
   • Texas Electronics TR-525M tipping bucket rain gauge
5. Air temperature, humidity, & vapor pressure deficit
   • Vaisala HMP155 probe
6. Soil temperature & water content
   • 3x Stevens HydraSense II probes buried at 5cm depth
7. Water level
   • 3x Hobo pressure transducers, replicate sensors within footprint

Validating some relationships between variables

A couple extra things: water chemistry of Kimages Creek (tidal creek running through footprint at US-RRC).

Our site is freshwater: usually < 0.2 ppt salinity year-round.

We have some exciting half-hourly phenocam data that will ultimately be processed and incorporated into future modeling efforts. Here are two example daily timeseries gifs from 2022 illustrating phenological change in our footprint over time.

April 29, 2022

August 30, 2022

Kimages Nitrate

Kimages chloride

Part 2: \(CO_2\) & \(CH_4\) Timeseries, June 6 - November 1, 2022

Part 3: Modeling NEE

Total NEE model:

## 
## Call:
## lm(formula = NEE ~ PPFD + Tair + VPD + TsoilMean + SWCMean + 
##     WaterLevelMean + WaterLevelDiff, data = rrc22_noNAs)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -42.256  -3.867  -0.651   2.551  76.692 
## 
## Coefficients:
##                  Estimate Std. Error t value Pr(>|t|)    
## (Intercept)     1.040e+02  1.587e+01   6.551 8.29e-11 ***
## PPFD           -1.580e-02  4.960e-04 -31.851  < 2e-16 ***
## Tair           -3.728e-01  1.217e-01  -3.065  0.00223 ** 
## VPD             2.095e-01  5.016e-02   4.176 3.17e-05 ***
## TsoilMean       2.667e-01  2.351e-01   1.134  0.25696    
## SWCMean        -1.560e+02  2.676e+01  -5.827 7.16e-09 ***
## WaterLevelMean  9.927e+00  4.722e+00   2.102  0.03575 *  
## WaterLevelDiff  1.610e+01  1.801e+01   0.894  0.37162    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 7.376 on 1256 degrees of freedom
## Multiple R-squared:  0.6899, Adjusted R-squared:  0.6882 
## F-statistic: 399.2 on 7 and 1256 DF,  p-value: < 2.2e-16

## Start:  AIC=5059.46
## NEE ~ PPFD + Tair + VPD + TsoilMean + SWCMean + WaterLevelMean + 
##     WaterLevelDiff
## 
##                  Df Sum of Sq    RSS    AIC
## - WaterLevelDiff  1        43  68374 5058.3
## - TsoilMean       1        70  68401 5058.8
## <none>                         68331 5059.5
## - WaterLevelMean  1       240  68571 5061.9
## - Tair            1       511  68842 5066.9
## - VPD             1       949  69280 5074.9
## - SWCMean         1      1847  70178 5091.2
## - PPFD            1     55192 123523 5805.8
## 
## Step:  AIC=5058.27
## NEE ~ PPFD + Tair + VPD + TsoilMean + SWCMean + WaterLevelMean
## 
##                  Df Sum of Sq    RSS    AIC
## - TsoilMean       1        87  68462 5057.9
## <none>                         68374 5058.3
## + WaterLevelDiff  1        43  68331 5059.5
## - WaterLevelMean  1       307  68681 5061.9
## - Tair            1       524  68899 5065.9
## - VPD             1       971  69346 5074.1
## - SWCMean         1      1853  70228 5090.1
## - PPFD            1     55483 123858 5807.3
## 
## Step:  AIC=5057.88
## NEE ~ PPFD + Tair + VPD + SWCMean + WaterLevelMean
## 
##                  Df Sum of Sq    RSS    AIC
## <none>                         68462 5057.9
## + TsoilMean       1        87  68374 5058.3
## + WaterLevelDiff  1        61  68401 5058.8
## - WaterLevelMean  1       231  68693 5060.1
## - Tair            1       505  68967 5065.2
## - VPD             1       893  69354 5072.3
## - SWCMean         1      1766  70228 5088.1
## - PPFD            1     86152 154614 6085.6

## Stepwise Model Path 
## Analysis of Deviance Table
## 
## Initial Model:
## NEE ~ PPFD + Tair + VPD + TsoilMean + SWCMean + WaterLevelMean + 
##     WaterLevelDiff
## 
## Final Model:
## NEE ~ PPFD + Tair + VPD + SWCMean + WaterLevelMean
## 
## 
##               Step Df Deviance Resid. Df Resid. Dev      AIC
## 1                                   1256   68331.00 5059.464
## 2 - WaterLevelDiff  1 43.45881      1257   68374.46 5058.268
## 3      - TsoilMean  1 87.47415      1258   68461.93 5057.884

## 
## Call:
## lm(formula = NEE ~ PPFD + Tair + VPD + SWCMean + WaterLevelMean, 
##     data = rrc22_noNAs)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -42.196  -3.962  -0.669   2.621  76.660 
## 
## Coefficients:
##                  Estimate Std. Error t value Pr(>|t|)    
## (Intercept)     1.039e+02  1.585e+01   6.556 8.04e-11 ***
## PPFD           -1.618e-02  4.066e-04 -39.788  < 2e-16 ***
## Tair           -2.733e-01  8.972e-02  -3.046  0.00237 ** 
## VPD             1.985e-01  4.901e-02   4.050 5.44e-05 ***
## SWCMean        -1.485e+02  2.607e+01  -5.697 1.52e-08 ***
## WaterLevelMean  8.880e+00  4.306e+00   2.062  0.03940 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 7.377 on 1258 degrees of freedom
## Multiple R-squared:  0.6893, Adjusted R-squared:  0.6881 
## F-statistic: 558.2 on 5 and 1258 DF,  p-value: < 2.2e-16

	Total NEE
Predictors	Estimates	Statistic	p
(Intercept)	103.917 (72.821 – 135.013)	6.556	<0.001
PPFD	-0.016 (-0.017 – -0.015)	-39.788	<0.001
Tair	-0.273 (-0.449 – -0.097)	-3.046	0.002
VPD	0.198 (0.102 – 0.295)	4.050	<0.001
SWCMean	-148.509 (-199.655 – -97.363)	-5.697	<0.001
WaterLevelMean	8.880 (0.432 – 17.328)	2.062	0.039
Observations	1264
R2 / R2 adjusted	0.689 / 0.688

Daytime NEE model:

## 
## Call:
## lm(formula = NEE ~ PPFD + Tair + VPD + TsoilMean + SWCMean + 
##     WaterLevelMean, data = daytime_data_noNAs)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -39.168  -3.538  -0.141   2.944  77.634 
## 
## Coefficients:
##                  Estimate Std. Error t value Pr(>|t|)    
## (Intercept)     1.111e+02  1.599e+01   6.948 7.45e-12 ***
## PPFD           -1.312e-02  4.953e-04 -26.498  < 2e-16 ***
## Tair           -6.715e-01  1.405e-01  -4.780 2.07e-06 ***
## VPD             3.426e-01  5.122e-02   6.689 4.12e-11 ***
## TsoilMean       4.469e-01  2.506e-01   1.783   0.0749 .  
## SWCMean        -1.690e+02  2.666e+01  -6.340 3.77e-10 ***
## WaterLevelMean  3.535e+00  5.991e+00   0.590   0.5553    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 6.37 on 834 degrees of freedom
## Multiple R-squared:  0.5931, Adjusted R-squared:  0.5902 
## F-statistic: 202.6 on 6 and 834 DF,  p-value: < 2.2e-16

## Start:  AIC=3121.25
## NEE ~ PPFD + Tair + VPD + TsoilMean + SWCMean + WaterLevelMean
## 
##                  Df Sum of Sq   RSS    AIC
## - WaterLevelMean  1      14.1 33851 3119.6
## <none>                        33837 3121.2
## - TsoilMean       1     129.0 33966 3122.4
## - Tair            1     927.1 34764 3142.0
## - SWCMean         1    1630.7 35467 3158.8
## - VPD             1    1815.2 35652 3163.2
## - PPFD            1   28485.9 62323 3632.9
## 
## Step:  AIC=3119.6
## NEE ~ PPFD + Tair + VPD + TsoilMean + SWCMean
## 
##                  Df Sum of Sq   RSS    AIC
## <none>                        33851 3119.6
## - TsoilMean       1     115.8 33967 3120.5
## + WaterLevelMean  1      14.1 33837 3121.2
## - Tair            1     934.7 34785 3140.5
## - SWCMean         1    1636.7 35487 3157.3
## - VPD             1    1842.4 35693 3162.2
## - PPFD            1   29010.0 62861 3638.1

## Stepwise Model Path 
## Analysis of Deviance Table
## 
## Initial Model:
## NEE ~ PPFD + Tair + VPD + TsoilMean + SWCMean + WaterLevelMean
## 
## Final Model:
## NEE ~ PPFD + Tair + VPD + TsoilMean + SWCMean
## 
## 
##               Step Df Deviance Resid. Df Resid. Dev      AIC
## 1                                    834   33836.57 3121.248
## 2 - WaterLevelMean  1 14.12522       835   33850.69 3119.599

## 
## Call:
## lm(formula = NEE ~ PPFD + Tair + VPD + TsoilMean + SWCMean, data = daytime_data_noNAs)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -38.983  -3.463  -0.136   2.929  77.674 
## 
## Coefficients:
##               Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  1.112e+02  1.598e+01   6.954 7.19e-12 ***
## PPFD        -1.316e-02  4.918e-04 -26.751  < 2e-16 ***
## Tair        -6.739e-01  1.404e-01  -4.802 1.86e-06 ***
## VPD          3.445e-01  5.110e-02   6.741 2.93e-11 ***
## TsoilMean    3.863e-01  2.285e-01   1.690   0.0913 .  
## SWCMean     -1.658e+02  2.610e+01  -6.354 3.45e-10 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 6.367 on 835 degrees of freedom
## Multiple R-squared:  0.593,  Adjusted R-squared:  0.5905 
## F-statistic: 243.3 on 5 and 835 DF,  p-value: < 2.2e-16

	Daytime NEE
Predictors	Estimates	Statistic	p
(Intercept)	111.151 (79.776 – 142.526)	6.954	<0.001
PPFD	-0.013 (-0.014 – -0.012)	-26.751	<0.001
Tair	-0.674 (-0.949 – -0.398)	-4.802	<0.001
VPD	0.345 (0.244 – 0.445)	6.741	<0.001
TsoilMean	0.386 (-0.062 – 0.835)	1.690	0.091
SWCMean	-165.849 (-217.082 – -114.617)	-6.354	<0.001
Observations	841
R2 / R2 adjusted	0.593 / 0.591

## [1] 0.461497

## [1] 0.02687054

## [1] 0.05161655

## [1] 0.003410281

## [1] 0.04612084

Nighttime NEE model:

## 
## Call:
## lm(formula = NEE ~ Tair + VPD + TsoilMean + SWCMean + WaterLevelMean, 
##     data = nighttime_data_noNAs)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -31.991  -4.148  -2.224   0.895  49.417 
## 
## Coefficients:
##                Estimate Std. Error t value Pr(>|t|)   
## (Intercept)      5.0666    37.1781   0.136  0.89167   
## Tair             0.5711     0.2169   2.633  0.00877 **
## VPD              0.1880     0.1235   1.522  0.12868   
## TsoilMean       -1.4271     0.4960  -2.877  0.00422 **
## SWCMean         40.0983    65.0718   0.616  0.53809   
## WaterLevelMean   1.3605     7.1179   0.191  0.84851   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 8.181 on 417 degrees of freedom
## Multiple R-squared:  0.04062,    Adjusted R-squared:  0.02912 
## F-statistic: 3.531 on 5 and 417 DF,  p-value: 0.003873

## Start:  AIC=1784.11
## NEE ~ Tair + VPD + TsoilMean + SWCMean + WaterLevelMean
## 
##                  Df Sum of Sq   RSS    AIC
## - WaterLevelMean  1      2.45 27913 1782.2
## - SWCMean         1     25.42 27936 1782.5
## <none>                        27911 1784.1
## - VPD             1    155.12 28066 1784.5
## - Tair            1    464.10 28375 1789.1
## - TsoilMean       1    554.11 28465 1790.4
## 
## Step:  AIC=1782.15
## NEE ~ Tair + VPD + TsoilMean + SWCMean
## 
##                  Df Sum of Sq   RSS    AIC
## - SWCMean         1     34.92 27948 1780.7
## <none>                        27913 1782.2
## - VPD             1    170.79 28084 1782.7
## + WaterLevelMean  1      2.45 27911 1784.1
## - Tair            1    467.72 28381 1787.2
## - TsoilMean       1    652.72 28566 1789.9
## 
## Step:  AIC=1780.68
## NEE ~ Tair + VPD + TsoilMean
## 
##                  Df Sum of Sq   RSS    AIC
## <none>                        27948 1780.7
## - VPD             1    136.34 28084 1780.7
## + SWCMean         1     34.92 27913 1782.2
## + WaterLevelMean  1     11.95 27936 1782.5
## - Tair            1    469.72 28418 1785.7
## - TsoilMean       1    620.06 28568 1788.0

## Stepwise Model Path 
## Analysis of Deviance Table
## 
## Initial Model:
## NEE ~ Tair + VPD + TsoilMean + SWCMean + WaterLevelMean
## 
## Final Model:
## NEE ~ Tair + VPD + TsoilMean
## 
## 
##               Step Df  Deviance Resid. Df Resid. Dev      AIC
## 1                                     417   27910.73 1784.114
## 2 - WaterLevelMean  1  2.445442       418   27913.17 1782.151
## 3        - SWCMean  1 34.915452       419   27948.09 1780.680

## 
## Call:
## lm(formula = NEE ~ Tair + VPD + TsoilMean, data = nighttime_data_noNAs)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -30.742  -4.065  -2.317   1.000  49.583 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  28.7753     8.3438   3.449  0.00062 ***
## Tair          0.5740     0.2163   2.654  0.00826 ** 
## VPD           0.1495     0.1046   1.430  0.15354    
## TsoilMean    -1.3619     0.4467  -3.049  0.00244 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 8.167 on 419 degrees of freedom
## Multiple R-squared:  0.03934,    Adjusted R-squared:  0.03246 
## F-statistic: 5.719 on 3 and 419 DF,  p-value: 0.0007639

	Nighttime NEE
Predictors	Estimates	Statistic	p
(Intercept)	28.775 (12.374 – 45.176)	3.449	0.001
Tair	0.574 (0.149 – 0.999)	2.654	0.008
VPD	0.150 (-0.056 – 0.355)	1.430	0.154
TsoilMean	-1.362 (-2.240 – -0.484)	-3.049	0.002
Observations	423
R2 / R2 adjusted	0.039 / 0.032

Plotting bivariate relationships for Daytime NEE

Plotting bivariate relationships for Nighttime NEE

Part 4: Modeling \(CH_4\)

## 
## Call:
## lm(formula = FCH4 ~ PPFD + Tair + VPD + TsoilMean + SWCMean + 
##     WaterLevelMean + WaterLevelDiff, data = rrc22_noNAs)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -0.54034 -0.07521 -0.00937  0.05726  0.93442 
## 
## Coefficients:
##                  Estimate Std. Error t value Pr(>|t|)    
## (Intercept)     3.535e+00  2.605e-01  13.573  < 2e-16 ***
## PPFD            1.479e-05  8.141e-06   1.817 0.069530 .  
## Tair            1.807e-02  1.997e-03   9.052  < 2e-16 ***
## VPD            -2.774e-04  8.234e-04  -0.337 0.736199    
## TsoilMean      -1.482e-02  3.859e-03  -3.839 0.000129 ***
## SWCMean        -5.473e+00  4.393e-01 -12.458  < 2e-16 ***
## WaterLevelMean  6.919e-01  7.751e-02   8.926  < 2e-16 ***
## WaterLevelDiff -8.686e-02  2.957e-01  -0.294 0.768988    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.1211 on 1256 degrees of freedom
## Multiple R-squared:  0.3226, Adjusted R-squared:  0.3188 
## F-statistic: 85.44 on 7 and 1256 DF,  p-value: < 2.2e-16

## Start:  AIC=-5329.6
## FCH4 ~ PPFD + Tair + VPD + TsoilMean + SWCMean + WaterLevelMean + 
##     WaterLevelDiff
## 
##                  Df Sum of Sq    RSS     AIC
## - WaterLevelDiff  1   0.00126 18.412 -5331.5
## - VPD             1   0.00166 18.412 -5331.5
## <none>                        18.410 -5329.6
## - PPFD            1   0.04837 18.459 -5328.3
## - TsoilMean       1   0.21608 18.627 -5316.9
## - WaterLevelMean  1   1.16791 19.578 -5253.9
## - Tair            1   1.20096 19.611 -5251.7
## - SWCMean         1   2.27513 20.686 -5184.3
## 
## Step:  AIC=-5331.51
## FCH4 ~ PPFD + Tair + VPD + TsoilMean + SWCMean + WaterLevelMean
## 
##                  Df Sum of Sq    RSS     AIC
## - VPD             1   0.00182 18.414 -5333.4
## <none>                        18.412 -5331.5
## - PPFD            1   0.04927 18.461 -5330.1
## + WaterLevelDiff  1   0.00126 18.410 -5329.6
## - TsoilMean       1   0.22481 18.637 -5318.2
## - Tair            1   1.20618 19.618 -5253.3
## - WaterLevelMean  1   1.21584 19.628 -5252.7
## - SWCMean         1   2.27427 20.686 -5186.3
## 
## Step:  AIC=-5333.39
## FCH4 ~ PPFD + Tair + TsoilMean + SWCMean + WaterLevelMean
## 
##                  Df Sum of Sq    RSS     AIC
## <none>                        18.414 -5333.4
## - PPFD            1   0.04775 18.461 -5332.1
## + VPD             1   0.00182 18.412 -5331.5
## + WaterLevelDiff  1   0.00142 18.412 -5331.5
## - TsoilMean       1   0.22626 18.640 -5320.0
## - WaterLevelMean  1   1.21575 19.629 -5254.6
## - Tair            1   1.77837 20.192 -5218.9
## - SWCMean         1   2.51719 20.931 -5173.4

## Stepwise Model Path 
## Analysis of Deviance Table
## 
## Initial Model:
## FCH4 ~ PPFD + Tair + VPD + TsoilMean + SWCMean + WaterLevelMean + 
##     WaterLevelDiff
## 
## Final Model:
## FCH4 ~ PPFD + Tair + TsoilMean + SWCMean + WaterLevelMean
## 
## 
##               Step Df    Deviance Resid. Df Resid. Dev       AIC
## 1                                      1256   18.41051 -5329.601
## 2 - WaterLevelDiff  1 0.001264936      1257   18.41177 -5331.514
## 3            - VPD  1 0.001816256      1258   18.41359 -5333.390

## 
## Call:
## lm(formula = FCH4 ~ PPFD + Tair + TsoilMean + SWCMean + WaterLevelMean, 
##     data = rrc22_noNAs)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -0.54042 -0.07554 -0.00943  0.05695  0.93398 
## 
## Coefficients:
##                  Estimate Std. Error t value Pr(>|t|)    
## (Intercept)     3.507e+00  2.448e-01  14.329  < 2e-16 ***
## PPFD            1.414e-05  7.830e-06   1.806   0.0711 .  
## Tair            1.768e-02  1.604e-03  11.023  < 2e-16 ***
## TsoilMean      -1.469e-02  3.737e-03  -3.932  8.9e-05 ***
## SWCMean        -5.420e+00  4.133e-01 -13.114  < 2e-16 ***
## WaterLevelMean  6.846e-01  7.512e-02   9.114  < 2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.121 on 1258 degrees of freedom
## Multiple R-squared:  0.3225, Adjusted R-squared:  0.3198 
## F-statistic: 119.7 on 5 and 1258 DF,  p-value: < 2.2e-16

	Total Methane
Predictors	Estimates	Statistic	p
(Intercept)	3.507 (3.027 – 3.988)	14.329	<0.001
PPFD	0.000 (-0.000 – 0.000)	1.806	0.071
Tair	0.018 (0.015 – 0.021)	11.023	<0.001
TsoilMean	-0.015 (-0.022 – -0.007)	-3.932	<0.001
SWCMean	-5.420 (-6.230 – -4.609)	-13.114	<0.001
WaterLevelMean	0.685 (0.537 – 0.832)	9.114	<0.001
Observations	1264
R2 / R2 adjusted	0.322 / 0.320

## 
## Call:
## lm(formula = FCH4 ~ PPFD + Tair + VPD + TsoilMean + SWCMean + 
##     WaterLevelMean + WaterLevelDiff, data = daytime_data_noNAs)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -0.51880 -0.06776 -0.00411  0.05507  0.46534 
## 
## Coefficients:
##                  Estimate Std. Error t value Pr(>|t|)    
## (Intercept)     3.628e+00  2.556e-01  14.193  < 2e-16 ***
## PPFD            3.887e-05  7.961e-06   4.883 1.25e-06 ***
## Tair            1.319e-02  2.244e-03   5.877 6.02e-09 ***
## VPD             1.322e-03  8.202e-04   1.611    0.108    
## TsoilMean       2.360e-03  4.007e-03   0.589    0.556    
## SWCMean        -6.293e+00  4.259e-01 -14.776  < 2e-16 ***
## WaterLevelMean  1.296e+00  9.575e-02  13.536  < 2e-16 ***
## WaterLevelDiff  7.063e-02  5.320e-01   0.133    0.894    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.1017 on 833 degrees of freedom
## Multiple R-squared:  0.454,  Adjusted R-squared:  0.4494 
## F-statistic: 98.95 on 7 and 833 DF,  p-value: < 2.2e-16

## Start:  AIC=-3836.22
## FCH4 ~ PPFD + Tair + VPD + TsoilMean + SWCMean + WaterLevelMean + 
##     WaterLevelDiff
## 
##                  Df Sum of Sq     RSS     AIC
## - WaterLevelDiff  1   0.00018  8.6201 -3838.2
## - TsoilMean       1   0.00359  8.6235 -3837.9
## <none>                         8.6199 -3836.2
## - VPD             1   0.02686  8.6467 -3835.6
## - PPFD            1   0.24670  8.8666 -3814.5
## - Tair            1   0.35747  8.9774 -3804.0
## - WaterLevelMean  1   1.89605 10.5159 -3671.0
## - SWCMean         1   2.25929 10.8792 -3642.5
## 
## Step:  AIC=-3838.2
## FCH4 ~ PPFD + Tair + VPD + TsoilMean + SWCMean + WaterLevelMean
## 
##                  Df Sum of Sq     RSS     AIC
## - TsoilMean       1   0.00367  8.6237 -3839.8
## <none>                         8.6201 -3838.2
## - VPD             1   0.02733  8.6474 -3837.5
## + WaterLevelDiff  1   0.00018  8.6199 -3836.2
## - PPFD            1   0.24840  8.8685 -3816.3
## - Tair            1   0.35730  8.9774 -3806.0
## - WaterLevelMean  1   1.90070 10.5208 -3672.6
## - SWCMean         1   2.25941 10.8795 -3644.4
## 
## Step:  AIC=-3839.84
## FCH4 ~ PPFD + Tair + VPD + SWCMean + WaterLevelMean
## 
##                  Df Sum of Sq     RSS     AIC
## <none>                         8.6237 -3839.8
## - VPD             1   0.02394  8.6477 -3839.5
## + TsoilMean       1   0.00367  8.6201 -3838.2
## + WaterLevelDiff  1   0.00027  8.6235 -3837.9
## - PPFD            1   0.28580  8.9095 -3814.4
## - Tair            1   0.74083  9.3646 -3772.5
## - WaterLevelMean  1   2.20274 10.8265 -3650.5
## - SWCMean         1   2.32193 10.9457 -3641.3

## Stepwise Model Path 
## Analysis of Deviance Table
## 
## Initial Model:
## FCH4 ~ PPFD + Tair + VPD + TsoilMean + SWCMean + WaterLevelMean + 
##     WaterLevelDiff
## 
## Final Model:
## FCH4 ~ PPFD + Tair + VPD + SWCMean + WaterLevelMean
## 
## 
##               Step Df     Deviance Resid. Df Resid. Dev       AIC
## 1                                        833   8.619885 -3836.217
## 2 - WaterLevelDiff  1 0.0001824028       834   8.620067 -3838.199
## 3      - TsoilMean  1 0.0036737584       835   8.623741 -3839.841

## 
## Call:
## lm(formula = FCH4 ~ PPFD + Tair + VPD + SWCMean + WaterLevelMean, 
##     data = daytime_data_noNAs)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -0.52231 -0.06696 -0.00474  0.05446  0.46146 
## 
## Coefficients:
##                  Estimate Std. Error t value Pr(>|t|)    
## (Intercept)     3.635e+00  2.547e-01  14.273  < 2e-16 ***
## PPFD            3.650e-05  6.938e-06   5.260 1.83e-07 ***
## Tair            1.408e-02  1.662e-03   8.469  < 2e-16 ***
## VPD             1.199e-03  7.873e-04   1.522    0.128    
## SWCMean        -6.239e+00  4.161e-01 -14.994  < 2e-16 ***
## WaterLevelMean  1.273e+00  8.719e-02  14.604  < 2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.1016 on 835 degrees of freedom
## Multiple R-squared:  0.4538, Adjusted R-squared:  0.4505 
## F-statistic: 138.7 on 5 and 835 DF,  p-value: < 2.2e-16

	Daytime Methane
Predictors	Estimates	Statistic	p
(Intercept)	3.63525 (3.13535 – 4.13516)	14.27334	<0.001
PPFD	0.00004 (0.00002 – 0.00005)	5.26045	<0.001
Tair	0.01408 (0.01082 – 0.01734)	8.46947	<0.001
VPD	0.00120 (-0.00035 – 0.00274)	1.52242	0.128
SWCMean	-6.23909 (-7.05582 – -5.42236)	-14.99409	<0.001
WaterLevelMean	1.27329 (1.10216 – 1.44442)	14.60419	<0.001
Observations	841
R2 / R2 adjusted	0.454 / 0.450

## [1] 0.03207744

## [1] 0.07911036

## [1] 0.002768067

## [1] 0.2121323

## [1] 0.2034589

## 
## Call:
## lm(formula = FCH4 ~ PPFD + Tair + VPD + TsoilMean + SWCMean + 
##     WaterLevelMean + WaterLevelDiff, data = nighttime_data_noNAs)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -0.48215 -0.08375 -0.02113  0.05192  0.89085 
## 
## Coefficients:
##                  Estimate Std. Error t value Pr(>|t|)    
## (Intercept)     2.2801276  0.6491326   3.513 0.000493 ***
## PPFD           -0.0008567  0.0015577  -0.550 0.582620    
## Tair            0.0301030  0.0037802   7.963 1.62e-14 ***
## VPD            -0.0013005  0.0021536  -0.604 0.546244    
## TsoilMean      -0.0504070  0.0087543  -5.758 1.66e-08 ***
## SWCMean        -2.3022192  1.1353063  -2.028 0.043215 *  
## WaterLevelMean  0.0212895  0.1311698   0.162 0.871145    
## WaterLevelDiff  0.4566001  0.4036559   1.131 0.258640    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.1423 on 415 degrees of freedom
## Multiple R-squared:  0.1815, Adjusted R-squared:  0.1677 
## F-statistic: 13.14 on 7 and 415 DF,  p-value: 2.541e-15

## Start:  AIC=-1641.86
## FCH4 ~ PPFD + Tair + VPD + TsoilMean + SWCMean + WaterLevelMean + 
##     WaterLevelDiff
## 
##                  Df Sum of Sq    RSS     AIC
## - WaterLevelMean  1   0.00053 8.3993 -1643.8
## - PPFD            1   0.00612 8.4049 -1643.5
## - VPD             1   0.00738 8.4061 -1643.5
## - WaterLevelDiff  1   0.02589 8.4246 -1642.6
## <none>                        8.3987 -1641.9
## - SWCMean         1   0.08322 8.4820 -1639.7
## - TsoilMean       1   0.67098 9.0697 -1611.3
## - Tair            1   1.28336 9.6821 -1583.7
## 
## Step:  AIC=-1643.83
## FCH4 ~ PPFD + Tair + VPD + TsoilMean + SWCMean + WaterLevelDiff
## 
##                  Df Sum of Sq    RSS     AIC
## - PPFD            1   0.00658 8.4059 -1645.5
## - VPD             1   0.00689 8.4062 -1645.5
## - WaterLevelDiff  1   0.03153 8.4308 -1644.2
## <none>                        8.3993 -1643.8
## + WaterLevelMean  1   0.00053 8.3987 -1641.9
## - SWCMean         1   0.08861 8.4879 -1641.4
## - TsoilMean       1   0.78970 9.1890 -1607.8
## - Tair            1   1.29138 9.6907 -1585.3
## 
## Step:  AIC=-1645.5
## FCH4 ~ Tair + VPD + TsoilMean + SWCMean + WaterLevelDiff
## 
##                  Df Sum of Sq    RSS     AIC
## - VPD             1   0.00754 8.4134 -1647.1
## - WaterLevelDiff  1   0.03097 8.4368 -1646.0
## <none>                        8.4059 -1645.5
## + PPFD            1   0.00658 8.3993 -1643.8
## + WaterLevelMean  1   0.00099 8.4049 -1643.5
## - SWCMean         1   0.09102 8.4969 -1643.0
## - TsoilMean       1   0.80035 9.2062 -1609.0
## - Tair            1   1.29269 9.6985 -1587.0
## 
## Step:  AIC=-1647.12
## FCH4 ~ Tair + TsoilMean + SWCMean + WaterLevelDiff
## 
##                  Df Sum of Sq    RSS     AIC
## - WaterLevelDiff  1   0.02823 8.4416 -1647.7
## <none>                        8.4134 -1647.1
## + VPD             1   0.00754 8.4059 -1645.5
## + PPFD            1   0.00723 8.4062 -1645.5
## + WaterLevelMean  1   0.00022 8.4132 -1645.1
## - SWCMean         1   0.08906 8.5025 -1644.7
## - TsoilMean       1   0.79324 9.2066 -1611.0
## - Tair            1   1.44717 9.8606 -1582.0
## 
## Step:  AIC=-1647.71
## FCH4 ~ Tair + TsoilMean + SWCMean
## 
##                  Df Sum of Sq    RSS     AIC
## <none>                        8.4416 -1647.7
## + WaterLevelDiff  1   0.02823 8.4134 -1647.1
## + PPFD            1   0.00653 8.4351 -1646.0
## + VPD             1   0.00480 8.4368 -1646.0
## + WaterLevelMean  1   0.00478 8.4369 -1645.9
## - SWCMean         1   0.08508 8.5267 -1645.5
## - TsoilMean       1   0.77923 9.2209 -1612.4
## - Tair            1   1.44190 9.8835 -1583.0

## Stepwise Model Path 
## Analysis of Deviance Table
## 
## Initial Model:
## FCH4 ~ PPFD + Tair + VPD + TsoilMean + SWCMean + WaterLevelMean + 
##     WaterLevelDiff
## 
## Final Model:
## FCH4 ~ Tair + TsoilMean + SWCMean
## 
## 
##               Step Df     Deviance Resid. Df Resid. Dev       AIC
## 1                                        415   8.398746 -1641.860
## 2 - WaterLevelMean  1 0.0005331243       416   8.399279 -1643.833
## 3           - PPFD  1 0.0065808106       417   8.405860 -1645.501
## 4            - VPD  1 0.0075393972       418   8.413399 -1647.122
## 5 - WaterLevelDiff  1 0.0282280012       419   8.441627 -1647.705

## 
## Call:
## lm(formula = FCH4 ~ Tair + TsoilMean + SWCMean, data = nighttime_data_noNAs)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -0.48105 -0.08594 -0.01962  0.05335  0.89275 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  2.041559   0.539259   3.786 0.000176 ***
## Tair         0.029176   0.003449   8.460 4.52e-16 ***
## TsoilMean   -0.050291   0.008087  -6.219 1.21e-09 ***
## SWCMean     -1.895712   0.922506  -2.055 0.040502 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.1419 on 419 degrees of freedom
## Multiple R-squared:  0.1773, Adjusted R-squared:  0.1714 
## F-statistic:  30.1 on 3 and 419 DF,  p-value: < 2.2e-16

	Nighttime Methane
Predictors	Estimates	Statistic	p
(Intercept)	2.04156 (0.98157 – 3.10155)	3.78586	<0.001
Tair	0.02918 (0.02240 – 0.03596)	8.45982	<0.001
TsoilMean	-0.05029 (-0.06619 – -0.03440)	-6.21907	<0.001
SWCMean	-1.89571 (-3.70903 – -0.08240)	-2.05496	0.041
Observations	423
R2 / R2 adjusted	0.177 / 0.171

Plotting bivariate relationships for Daytime Methane Efflux

Note that the datasets required for AIC necessarily omitted all rows with any missing values, vastly truncating the total data set for that modeling exercise. So, some relationships plotted above would look different if plotted with all available data, rather than the AIC-friendly subsetted data.