Overall TP vs pCHL relationship by groups
GAMs by Lake group
Group1
##
## Family: gaussian
## Link function: identity
##
## Formula:
## logpCHL ~ s(logTP) + s(logTN) + s(logNTU) + s(logNVSS) + s(Year)
##
## Parametric coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.94774 0.01828 161.2 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Approximate significance of smooth terms:
## edf Ref.df F p-value
## s(logTP) 2.640 3.367 61.762 < 2e-16 ***
## s(logTN) 2.548 3.285 17.284 1.65e-11 ***
## s(logNTU) 3.222 4.067 164.494 < 2e-16 ***
## s(logNVSS) 4.813 5.972 2.664 0.016586 *
## s(Year) 7.386 8.371 3.594 0.000255 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## R-sq.(adj) = 0.514 Deviance explained = 52.4%
## GCV = 0.33798 Scale est. = 0.33059 n = 989
Group 2
##
## Family: gaussian
## Link function: identity
##
## Formula:
## logpCHL ~ s(logTP) + s(logTN) + s(logNTU) + s(logNVSS) + s(Year)
##
## Parametric coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.73056 0.01848 147.8 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Approximate significance of smooth terms:
## edf Ref.df F p-value
## s(logTP) 4.910 6.045 45.091 < 2e-16 ***
## s(logTN) 4.251 5.293 26.545 < 2e-16 ***
## s(logNTU) 6.588 7.764 7.403 4.17e-09 ***
## s(logNVSS) 2.230 2.841 2.533 0.0617 .
## s(Year) 3.859 4.672 3.012 0.0120 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## R-sq.(adj) = 0.826 Deviance explained = 83.3%
## GCV = 0.19802 Scale est. = 0.18988 n = 556
Group 3
##
## Family: gaussian
## Link function: identity
##
## Formula:
## logpCHL ~ s(logTP) + s(logTN) + s(logNTU) + s(logNVSS) + s(Year)
##
## Parametric coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.59800 0.02616 22.86 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Approximate significance of smooth terms:
## edf Ref.df F p-value
## s(logTP) 1.910 2.434 46.460 < 2e-16 ***
## s(logTN) 4.483 5.538 2.771 0.015908 *
## s(logNTU) 5.501 6.654 3.423 0.002077 **
## s(logNVSS) 1.000 1.000 1.831 0.176957
## s(Year) 1.000 1.000 15.195 0.000117 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## R-sq.(adj) = 0.385 Deviance explained = 41.1%
## GCV = 0.24138 Scale est. = 0.23071 n = 337
Group 4
##
## Family: gaussian
## Link function: identity
##
## Formula:
## logpCHL ~ s(logTP) + s(logTN) + s(logNTU) + s(logNVSS) + s(Year)
##
## Parametric coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.56980 0.01124 228.7 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Approximate significance of smooth terms:
## edf Ref.df F p-value
## s(logTP) 3.996 4.988 200.810 < 2e-16 ***
## s(logTN) 4.184 5.225 27.475 < 2e-16 ***
## s(logNTU) 5.428 6.611 42.442 < 2e-16 ***
## s(logNVSS) 1.000 1.001 0.331 0.565
## s(Year) 7.512 8.463 4.055 5.96e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## R-sq.(adj) = 0.809 Deviance explained = 81.2%
## GCV = 0.21253 Scale est. = 0.20957 n = 1660
Group 5
##
## Family: gaussian
## Link function: identity
##
## Formula:
## logpCHL ~ s(logTP) + s(logTN) + s(logNTU) + s(logNVSS) + s(Year)
##
## Parametric coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.08564 0.02268 136.1 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Approximate significance of smooth terms:
## edf Ref.df F p-value
## s(logTP) 4.188 5.188 36.787 <2e-16 ***
## s(logTN) 3.661 4.652 23.253 <2e-16 ***
## s(logNTU) 8.534 8.935 24.018 <2e-16 ***
## s(logNVSS) 3.464 4.375 1.799 0.1256
## s(Year) 1.000 1.000 4.760 0.0296 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## R-sq.(adj) = 0.704 Deviance explained = 71.7%
## GCV = 0.26378 Scale est. = 0.25201 n = 490
Merge Groups 2 and 4 and Groups 1 and 5
Plot by groups
GAMS with TP, TN and NTU with month interaction
##
## Family: gaussian
## Link function: identity
##
## Formula:
## logpCHL ~ s(logTP, factor(Month), bs = "fs") + s(logTN, factor(Month),
## bs = "fs") + s(logNTU, factor(Month), bs = "fs")
##
## Parametric coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.7617 0.3789 9.929 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Approximate significance of smooth terms:
## edf Ref.df F p-value
## s(logTP,factor(Month)) 15.59 39 13.37 <2e-16 ***
## s(logTN,factor(Month)) 11.14 36 5.36 <2e-16 ***
## s(logNTU,factor(Month)) 12.77 39 21.45 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## R-sq.(adj) = 0.61 Deviance explained = 62.1%
## GCV = 0.29703 Scale est. = 0.2889 n = 1479
##
## Family: gaussian
## Link function: identity
##
## Formula:
## logpCHL ~ s(logTP, factor(Month), bs = "fs") + s(logTN, factor(Month),
## bs = "fs") + s(logNTU, factor(Month), bs = "fs")
##
## Parametric coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.862 9.602 0.298 0.766
##
## Approximate significance of smooth terms:
## edf Ref.df F p-value
## s(logTP,factor(Month)) 15.97 39 38.598 <2e-16 ***
## s(logTN,factor(Month)) 16.46 39 7.210 <2e-16 ***
## s(logNTU,factor(Month)) 13.95 39 5.858 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## R-sq.(adj) = 0.812 Deviance explained = 81.6%
## GCV = 0.21156 Scale est. = 0.20703 n = 2216
##
## Family: gaussian
## Link function: identity
##
## Formula:
## logpCHL ~ s(logTP, factor(Month), bs = "fs") + s(logTN, factor(Month),
## bs = "fs") + s(logNTU, factor(Month), bs = "fs")
##
## Parametric coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.6096 0.1555 3.92 0.000108 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Approximate significance of smooth terms:
## edf Ref.df F p-value
## s(logTP,factor(Month)) 3.856 28 4.739 < 2e-16 ***
## s(logTN,factor(Month)) 8.651 36 1.049 3.95e-06 ***
## s(logNTU,factor(Month)) 2.423 27 0.335 0.00693 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## R-sq.(adj) = 0.376 Deviance explained = 40.3%
## GCV = 0.24604 Scale est. = 0.23441 n = 337
Groups alternative 2
4 groups merge 2 and 4 only.
Plot by groups and Lakes
##
## Family: gaussian
## Link function: identity
##
## Formula:
## logpCHL ~ s(logTP, factor(Month), bs = "fs") + s(logTN, factor(Month),
## bs = "fs") + s(logNTU, factor(Month), bs = "fs")
##
## Parametric coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.7162 0.3158 11.77 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Approximate significance of smooth terms:
## edf Ref.df F p-value
## s(logTP,factor(Month)) 11.39 39 7.388 <2e-16 ***
## s(logTN,factor(Month)) 11.68 37 2.606 <2e-16 ***
## s(logNTU,factor(Month)) 10.22 39 13.802 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## R-sq.(adj) = 0.544 Deviance explained = 55.9%
## GCV = 0.32163 Scale est. = 0.31047 n = 989
Group2
##
## Family: gaussian
## Link function: identity
##
## Formula:
## logpCHL ~ s(logTP, factor(Month), bs = "fs") + s(logTN, factor(Month),
## bs = "fs") + s(logNTU, factor(Month), bs = "fs")
##
## Parametric coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.862 9.602 0.298 0.766
##
## Approximate significance of smooth terms:
## edf Ref.df F p-value
## s(logTP,factor(Month)) 15.97 39 38.598 <2e-16 ***
## s(logTN,factor(Month)) 16.46 39 7.210 <2e-16 ***
## s(logNTU,factor(Month)) 13.95 39 5.858 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## R-sq.(adj) = 0.812 Deviance explained = 81.6%
## GCV = 0.21156 Scale est. = 0.20703 n = 2216
Group3
##
## Family: gaussian
## Link function: identity
##
## Formula:
## logpCHL ~ s(logTP, factor(Month), bs = "fs") + s(logTN, factor(Month),
## bs = "fs") + s(logNTU, factor(Month), bs = "fs")
##
## Parametric coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.6096 0.1555 3.92 0.000108 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Approximate significance of smooth terms:
## edf Ref.df F p-value
## s(logTP,factor(Month)) 3.856 28 4.739 < 2e-16 ***
## s(logTN,factor(Month)) 8.651 36 1.049 3.95e-06 ***
## s(logNTU,factor(Month)) 2.423 27 0.335 0.00693 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## R-sq.(adj) = 0.376 Deviance explained = 40.3%
## GCV = 0.24604 Scale est. = 0.23441 n = 337
Group4
##
## Family: gaussian
## Link function: identity
##
## Formula:
## logpCHL ~ s(logTP, factor(Month), bs = "fs") + s(logTN, factor(Month),
## bs = "fs") + s(logNTU, factor(Month), bs = "fs")
##
## Parametric coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.255 1.241 3.428 0.000666 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Approximate significance of smooth terms:
## edf Ref.df F p-value
## s(logTP,factor(Month)) 27.61 39 7.561 <2e-16 ***
## s(logTN,factor(Month)) 10.57 37 3.359 <2e-16 ***
## s(logNTU,factor(Month)) 21.75 37 4.269 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## R-sq.(adj) = 0.764 Deviance explained = 79.3%
## GCV = 0.23013 Scale est. = 0.20151 n = 490
