AGC Model
- \(n\) train: 7757
- \(n\) test: 1918
- Predictors: live_agb, dem, slope, aspect, twi, tmin, tmax, precip,
chm
Random Forest Model - Test
Set 1:1
Params
$formula
agc ~ .
$num.trees
[1] 500
$mtry
[1] 5
$min.node.size
[1] 1
$replace
[1] FALSE
$sample.fraction
[1] 1
Stepwise Linear Model -
Test Set 1:1
Params
Call:
lm(formula = agc ~ live_agb + dem + slope + twi + tmin + tmax +
precip + chm, data = training_agc)
Residuals:
Min 1Q Median 3Q Max
-9.638e-04 -2.886e-04 4.513e-05 2.294e-04 1.975e-03
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 5.615e-04 9.568e-05 5.868e+00 4.58e-09 ***
live_agb 5.000e-01 5.688e-08 8.790e+06 < 2e-16 ***
dem -6.300e-08 3.197e-08 -1.971e+00 0.04881 *
slope 7.138e-06 1.189e-06 6.004e+00 2.02e-09 ***
twi 3.616e-05 5.442e-06 6.645e+00 3.24e-11 ***
tmin -1.226e-05 4.533e-06 -2.705e+00 0.00684 **
tmax -3.820e-05 5.770e-06 -6.621e+00 3.81e-11 ***
precip -6.994e-08 3.053e-08 -2.291e+00 0.02200 *
chm 5.977e-06 7.895e-07 7.571e+00 4.15e-14 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 0.0003334 on 7748 degrees of freedom
Multiple R-squared: 1, Adjusted R-squared: 1
F-statistic: 1.124e+13 on 8 and 7748 DF, p-value: < 2.2e-16
Test Set Accuracy
|
|
RF
|
LM
|
|
RMSE
|
0.330
|
0
|
|
MAE
|
0.172
|
0
|
|
MBE
|
0.011
|
0
|
|
R2
|
1.000
|
1
|
BGC Model
- \(n\) train: 7757
- \(n\) test: 1918
- Predictors: live_agb, dem, slope, aspect, twi, tmin, tmax, precip,
chm
Random Forest Model - Test
Set 1:1
Params
$formula
bgc ~ .
$num.trees
[1] 450
$mtry
[1] 5
$min.node.size
[1] 1
$replace
[1] FALSE
$sample.fraction
[1] 0.75
Stepwise Linear Model -
Test Set 1:1
Params
Call:
lm(formula = bgc ~ live_agb + dem + aspect + twi + tmin + precip +
chm, data = training_bgc)
Residuals:
Min 1Q Median 3Q Max
-1.2472 -0.2897 -0.1161 0.1433 3.6464
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 2.914e-01 6.876e-02 4.237 2.29e-05 ***
live_agb 9.904e-02 8.502e-05 1164.859 < 2e-16 ***
dem -3.192e-04 4.551e-05 -7.014 2.51e-12 ***
aspect 4.392e-04 6.312e-05 6.957 3.76e-12 ***
twi 2.825e-02 5.352e-03 5.279 1.34e-07 ***
tmin -8.773e-02 4.497e-03 -19.510 < 2e-16 ***
precip -1.163e-04 4.574e-05 -2.542 0.011 *
chm 5.678e-03 1.183e-03 4.798 1.63e-06 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 0.5014 on 7749 degrees of freedom
Multiple R-squared: 0.9951, Adjusted R-squared: 0.9951
F-statistic: 2.234e+05 on 7 and 7749 DF, p-value: < 2.2e-16
Test Set Accuracy
|
|
RF
|
LM
|
|
RMSE
|
0.410
|
0.474
|
|
MAE
|
0.291
|
0.336
|
|
MBE
|
0.034
|
0.008
|
|
R2
|
0.996
|
0.995
|
Deadwood C Model
- Outliers >= [(1.5 * IQR) + 0.75th percentile] of deadwood
measurements filtered: 25
- Outlier threshold = [(1.5 * IQR) + 0.75th percentile] = 22.479
- \(n\) train: 484
- \(n\) test: 115
- Predictors: live_agb, dem, slope, aspect, twi, tmin, tmax, precip,
chm
Random Forest Model - Test
Set 1:1
Params
$formula
deadwood ~ .
$num.trees
[1] 80
$mtry
[1] 1
$min.node.size
[1] 4
$replace
[1] TRUE
$sample.fraction
[1] 0.5
Stepwise Linear Model -
Test Set 1:1
Params
Call:
lm(formula = deadwood ~ live_agb + tmin + tmax, data = training_deadwood)
Residuals:
Min 1Q Median 3Q Max
-8.659 -3.499 -1.246 2.482 16.310
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 10.514464 3.700497 2.841 0.00468 **
live_agb 0.020798 0.003539 5.877 7.82e-09 ***
tmin -0.734992 0.259211 -2.835 0.00477 **
tmax -0.448465 0.315681 -1.421 0.15607
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 4.839 on 480 degrees of freedom
Multiple R-squared: 0.1858, Adjusted R-squared: 0.1807
F-statistic: 36.52 on 3 and 480 DF, p-value: < 2.2e-16
Test Set Accuracy
|
|
RF
|
LM
|
|
RMSE
|
4.031
|
4.463
|
|
MAE
|
3.153
|
3.497
|
|
MBE
|
-0.161
|
-0.300
|
|
R2
|
0.413
|
0.280
|
Litter C models
- \(n\) train: 127
- \(n\) test: 24
- Predictors: live_agb, dem, slope, aspect, twi, tmin, tmax, precip,
chm
Random Forest
Model - All Data (Train + Test) 1:1
Params
$formula
litter ~ .
$num.trees
[1] 500
$mtry
[1] 4
$min.node.size
[1] 1
$replace
[1] TRUE
$sample.fraction
[1] 0.95
Stepwise Linear
Model - All Data (Train + Test) 1:1
Params
Call:
lm(formula = litter ~ live_agb + twi + tmax + precip, data = training_litter)
Residuals:
Min 1Q Median 3Q Max
-13.332 -5.178 -2.687 3.239 45.603
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.020337 12.654842 -0.002 0.99872
live_agb 0.036272 0.012602 2.878 0.00472 **
twi 1.446577 0.685519 2.110 0.03688 *
tmax -1.858113 0.571105 -3.254 0.00147 **
precip 0.020024 0.006318 3.169 0.00193 **
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 8.43 on 122 degrees of freedom
Multiple R-squared: 0.2487, Adjusted R-squared: 0.2241
F-statistic: 10.1 on 4 and 122 DF, p-value: 4.287e-07
All Data (Train + Test)
Accuracy
|
|
RF
|
LM
|
|
RMSE
|
5.605
|
8.812
|
|
MAE
|
3.067
|
6.076
|
|
MBE
|
-0.488
|
-0.580
|
|
R2
|
0.703
|
0.266
|
Soil C model
- \(n\) train: 70
- \(n\) test: 17
- Predictors: live_agb, dem, slope, aspect, twi, tmin, tmax, precip,
chm
Random Forest
Model - All Data (Train + Test) 1:1
Params
$formula
soil ~ .
$num.trees
[1] 20
$mtry
[1] 2
$min.node.size
[1] 5
$replace
[1] TRUE
$sample.fraction
[1] 0.25
Stepwise
Linear Model - All Data (Train + Test) 1:1
Params
Call:
lm(formula = soil ~ live_agb + slope + tmax, data = training_soil)
Residuals:
Min 1Q Median 3Q Max
-71.29 -31.61 -15.87 12.38 404.58
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 255.2768 82.0059 3.113 0.00274 **
live_agb 0.2868 0.1236 2.320 0.02345 *
slope -2.9930 1.6348 -1.831 0.07164 .
tmax -13.2780 5.8765 -2.260 0.02716 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 66.21 on 66 degrees of freedom
Multiple R-squared: 0.1546, Adjusted R-squared: 0.1162
F-statistic: 4.024 on 3 and 66 DF, p-value: 0.01085
All Data (Train + Test)
Accuracy
|
|
RF
|
LM
|
|
RMSE
|
56.546
|
59.072
|
|
MAE
|
37.371
|
35.443
|
|
MBE
|
13.100
|
0.280
|
|
R2
|
0.218
|
0.146
|
Corrections
If you see mistakes or want to suggest changes, please create an issue on the source repository.
Citation
For attribution, please cite this work as
Johnson (2022, June 24). CAFRI Labs: Carbon Conversion - Take 2: Plot Measurements. Retrieved from https://cafri-labs.github.io/acceptable-growing-stock/posts/carbon-conversion-take-2-plot-measurements/
BibTeX citation
@misc{johnson2022carbon,
author = {Johnson, Lucas},
title = {CAFRI Labs: Carbon Conversion - Take 2: Plot Measurements},
url = {https://cafri-labs.github.io/acceptable-growing-stock/posts/carbon-conversion-take-2-plot-measurements/},
year = {2022}
}
