3.3.2. Prediction using alternate model
Table 3.
Standardized direct effects of predictor variables on predicted variables. Abbreviation: ODN = other dissolved nutrients.Env. flowPhytoplanktonTSSNitrogen0.34(0.13)[1.81]− 0.57(0.22)[2.59]0.07(0.08)[0.82]Path coefficientStandard errort-ValueODN0.21(0.09)[2.33]0.16(0.097)[1.63]0.19(0.08)[2.37]Path coefficientStandard errort-ValueFull-size tableTable optionsView in workspaceDownload as CSV
The correlation matrix using the ML method showed that EQKLISEEDL observed and predicted values were closely related. The five highest residuals between observed–predicted logNO2 + 3–logTEMP, logNH3–logTSS, logNH3–logSECCHI, logSAL–logPO4, and logTEMP–logSAL were 0.20, 0.13, − 0.13, − 0.11, and 0.09 respectively (Table 4).
Table 4.
4. Discussion
4.1. Distribution of inorganic nutrients
The three estuaries lie in a gradient of decreasing rainfall and thus decreasing freshwater inflow. The Lavaca–Colorado estuary has more rainfall than the Nueces and Guadalupe Estuaries (Table 1). The annual average surface inflow data from the Texas Development Water Development Board showed endoskeleton the Lavaca–Colorado Estuary has more annual inflow than the Guadalupe and Nueces Estuaries (Fig. 7). Though the Lavaca–Colorado Estuary has more inflow than the Guadalupe Estuary, the surface area of the former is almost double than that of the latter. Thus, the transported nutrient signals are more dilute in the Lavaca–Colorado Estuary than in the Guadalupe Estuary. This is the likely reason for the lower inorganic nutrient concentration in the Lavaca–Colorado Estuary compared to the Guadalupe Estuary.
Table 3.
Standardized direct effects of predictor variables on predicted variables. Abbreviation: ODN = other dissolved nutrients.Env. flowPhytoplanktonTSSNitrogen0.34(0.13)[1.81]− 0.57(0.22)[2.59]0.07(0.08)[0.82]Path coefficientStandard errort-ValueODN0.21(0.09)[2.33]0.16(0.097)[1.63]0.19(0.08)[2.37]Path coefficientStandard errort-ValueFull-size tableTable optionsView in workspaceDownload as CSV
The correlation matrix using the ML method showed that EQKLISEEDL observed and predicted values were closely related. The five highest residuals between observed–predicted logNO2 + 3–logTEMP, logNH3–logTSS, logNH3–logSECCHI, logSAL–logPO4, and logTEMP–logSAL were 0.20, 0.13, − 0.13, − 0.11, and 0.09 respectively (Table 4).
Table 4.
4. Discussion
4.1. Distribution of inorganic nutrients
The three estuaries lie in a gradient of decreasing rainfall and thus decreasing freshwater inflow. The Lavaca–Colorado estuary has more rainfall than the Nueces and Guadalupe Estuaries (Table 1). The annual average surface inflow data from the Texas Development Water Development Board showed endoskeleton the Lavaca–Colorado Estuary has more annual inflow than the Guadalupe and Nueces Estuaries (Fig. 7). Though the Lavaca–Colorado Estuary has more inflow than the Guadalupe Estuary, the surface area of the former is almost double than that of the latter. Thus, the transported nutrient signals are more dilute in the Lavaca–Colorado Estuary than in the Guadalupe Estuary. This is the likely reason for the lower inorganic nutrient concentration in the Lavaca–Colorado Estuary compared to the Guadalupe Estuary.