Using the STARS model to evaluate the effects of the proposed action for the reinitiation of consultation on the coordinated long-term operation of the Central Valley and State Water Project

Abstract

In 2016, the U.S. Bureau of Reclamation (USBR) and California Department of Water Resources requested a reinitiation of consultation under Section 7 of the Endangered Species Act on the coordinated long-term operations of the Central Valley and State Water Projects. This resulted in a Biological Assessment released by USBR in 2019. In its analysis of the Biological Assessment for its Biological Opinion on the proposed action, the National Marine Fisheries Service (NMFS) requested assistance from the U.S. Geological Survey to describe the effect of the proposed action on juvenile Chinook salmon ( Oncorhynchus tshawytscha ) populations migrating through the Sacramento-San Joaquin River Delta (henceforth called “the Delta”). Therefore, in this report we analyzed an 82-year time series of simulated river flows and Delta Cross Channel (DCC) gate operations under two scenarios constructed for the Biological Assessment: the proposed-action (PA) scenario and the continuing-operations scenario (COS). To evaluate the proposed action, we used the STARS model ( S urvival, T ravel time, A nd R outing S imulation model), a stochastic, individual-based simulation model designed to predict survival of a cohort of fish that experiences variable daily river flows as the fish migrate through the Delta. The STARS model uses parameter estimates from a Bayesian mark-recapture model that jointly estimates travel time and survival in eight discrete reaches of the Delta and migration routing at two key river junctions. By applying the STARS model to the two 82-year scenarios, we found that the proposed action had negative effects on survival, travel time, and routing in October–December but positive effects in April–June. In October–December, there was a high probability that survival in the PA scenario was less than that in the COS, and that travel time and routing to the Interior Delta for the PA scenario was greater than that for the COS. The magnitude of the difference in survival between scenarios was larger in some years than in others. For example, we quantified that survival under the PA scenario was 10 percent lower than under the COS in 25 percent of the water years from October through December. During this period, inflow to the Delta tended to be lower under the PA scenario, and the DCC gate was open more frequently under the PA scenario than during the COS. Lower inflow reduces survival, and more frequent operation of the DCC gate 1) increases the proportion of fish entering the Interior Delta, where survival is low, and thus 2) reduces survival in the Sacramento River in reaches downstream of the DCC. In contrast, during the period April–June, survival was higher, travel times were lower, and routing to the Interior Delta was lower under the PA scenario relative to the COS, although the magnitude of the increase in survival was relatively small in most years (less than a 3-percent difference in survival). This difference between scenarios was driven by higher river flows in some years under the PA scenario relative to the COS. Overall, the differences in survival, travel time, and routing distance between the two operational scenarios were primarily driven by the timing and magnitude of the annual high river flows.