2025 Orange County Water Demand Projection Model
Single-Family
Multifamily
CII
Irrigation
0
-0.05
-0.1
-0.15
-0.2
-0.25
-0.3
1
2
3
4
5
6
7
8
9 10 11 12
Calendar Month
Figure 3-6: Estimated Seasonal Price Elasticities by Demand Sector for Orange County Water Agencies
Even though the model coefficient for price is equal for each agency across all four sectors, the actual water rate data each agency provided is in the panel dataset, so the data by which this price coefficient is multiplied differs. Additionally, agencies adopt different block structures (different prices for the 10 th or the 20 th CCF) over time. Both the spatial and temporal variability in water rates allows us to estimate these elasticities, even though the coefficients are applied at the regional level. 3.8 Model Development Summary The econometric regressions show strong performance in explaining historical patterns of consumption over the 15 years from 2010 to 2024, including two major droughts and the COVID pandemic. Most of the regressions had R-squared values of 0.80 or greater. None of the regressions demonstrated a large, consistent bias. Based on this analysis, the estimated regression equations reflect a suitable basis for forecasting. The overall model approach allows for demand forecast scenario analysis based on varying assumptions of future conditions. Several forecast scenarios may be explored, including climate change-adjusted weather, alternate assumptions around the timing and magnitude of drought recovery, alternate assumptions around urban development, and/or different assumptions around future economic conditions. For any of these future scenarios, the model coefficients presented in this section should be maintained as they reflect the best-fitted estimates of the causal relationships between external socioeconomic conditions and historical water demand, given the available modeling data. Model scenarios can also be developed to address uncertainties in future explanatory variables, such as housing/job growth and density. Future inputs in these scenarios could be conducted as a sensitivity analysis or be driven by alternate growth projections. On a regular basis, overall model performance
3-15
Appendix G - 52
Powered by FlippingBook