2025 Orange County Water Demand Projection Model
Table 3-4: Multifamily Regression Variables and Coefficients
Explanatory Variable
General Directional Influence
Model Coefficient Range
PPH
+
0.5 (restricted)*
Housing Density
-
-0.33 0.79
MF Units per Account
+
Marginal Price (per 20 CCF) Departure from normal monthly maximum temperature Departure from normal monthly maximum temperature (lagged 1 month) Departure from normal monthly precipitation Departure from normal monthly precipitation (lagged 1 month)
-
-0.035
+
0 to 1.07
+
0 to 1.46
-
0 to -0.04
-
0 to -0.06
Residual Trend COVID Indicator
-
-0.007 to 0.0017
+
-0.05 to 0.14 -0.8 to 0.32
State Requested Percent Restriction
- -
State Drought Declaration -0.16 to -0.04 *The persons per household coefficient was set to 0.5, which was lower than the initial econometric estimate (which was >1). This change avoided instabilities, with little effect on the other coefficient estimates.
The multifamily use per account is less influenced by water price than single-family (lower coefficient).
3.4.2
Historical Model Performance
Visual inspection of the time series plots and review of the model fit parameters showed good model performance, with most fits exceeding an R 2 of 80%. Figure 3-3 s hows the performance of the multifamily regression for an example agency.
3-10
Appendix G - 47
Powered by FlippingBook