This analysis was used to inform the San Gabriel River adaptive management plan for flow management in the urbanized San Gabriel River, California, USA where modification of discharge of treated wastewater is desired to promote water reuse, without adversely affecting habitat and sensitive species reliant on the flows.
The objectives of the AMP are to ensure continuation of the pre-Project conditions (overall quality and quantity) of the habitat influenced by treatment plant discharges.
To accomplish this, it is important to understand: • Relationships between changes in WRP discharge and habitat quality • Sensitivity of the selected stressor s to changes in SGR river flow • Relative influence of changes in WRP discharge vs. other stressor variables on habitat • Intensity and duration of monitoring necessary to detect habitat changes associated with flow modifications
Note that there were many iterations during the model development, which is contained in the following scripts. Final models for both stressor indicators as in Irving et al xxx is located in scripts 5, 5a, 6 & 6a
data needed input_data/RawData_SGR.csv input_data/Daily_Data_Dec2022.csv
Formats monitoring data - position, substrate season and variable names Assigns groups to gages and effluent outfalls Plots data to visualise Formats climate data from PRISM - 34.0283_-118.0331 (https://prism.oregonstate.edu/) calculates monthly values from hydrology and climate data Adds in distance from outfall
Output is 00_bio_Q_matched_groups_distance.RData
All scripts are experimental aligning with questions and comments from Habitat Management Committee Can be ignored
Data needed 00_bio_Q_matched_groups_distance.RData
Stem Water Potential on observed values -
Random Forest (RF) - variable importance Mixed Model (MM) - significance of variables Predictive Model (PM) - probability of stress Power analysis (PA) - duration of monitoring
RF need import of external functions - Code/Functions/Functions.R RF cross validation at end of script (takes a long time to run) RF & MM run with and without damaged trees PM run on dry years and overall - names delta models in code PM Final figure is 05_probability_of_stress_overall_current_delta_range_ALL.jpg
Data needed 00_bio_Q_matched_groups_distance.RData
Stem Water Potential on relative values -
Random Forest (RF) - variable importance Mixed Model (MM) - significance of variables
RF need import of external functions - Code/Functions/Functions.R RF cross validation at end of script (takes a long time to run) RF & MM run with and without damaged trees
Data needed 00_bio_Q_matched_groups_distance.RData
Stem Water Potential on observed values -
Random Forest (RF) - variable importance Mixed Model (MM) - significance of variables Predictive Model (PM) - probability of stress Power analysis (PA) - duration of monitoring
RF need import of external functions - Code/Functions/Functions.R RF cross validation at end of script (takes a long time to run) RF & MM run with and without damaged trees PM run on dry years and overall - names delta models in code PM Final figure is 06_CV_probability_of_stress_overall_current_delta_range.jpg
Data needed 00_bio_Q_matched_groups_distance.RData
Stem Water Potential on relative values -
Random Forest (RF) - variable importance Mixed Model (MM) - significance of variables
RF need import of external functions - Code/Functions/Functions.R RF cross validation at end of script (takes a long time to run) RF & MM run with and without damaged trees
Sensitivity of proportional flow to group 5 for SWP relative im portance from RF tested main plot - 07_relative_importance_SWP_comparison_plot.jpg
Sensitivity of proportional flow to group 5 for CV relative importance from RF tested main plot - 08_relative_importance_CV_comparison_plot.jpg
Figures for SCCWRP symposium slide deck For animation purposes
Citations for packages used