UC Davis

Climate vulnerability assessments (CVA) lack a standardized approach to including long-term monitoringdata in sensitivity scoring. In this study, a scoring mechanism was produced to evaluate long-term monitoring data for sensitivity of wetland fish to water quality attributes. Suisun Marsh Fish Study data from 2011-2023 was used to produce a scoring mechanism for the response of three fish species (longfin smelt (Spirinchus thaleicthys), prickly sculpin (Cottus asper), and threespine stickleback (Gasterosteus aculeatus)) to water temperature (°C), salinity (ppt), turbidity (cm), and dissolved oxygen (mg/L). The relationships between these fish and the water quality of Suisun Marsh were assessed using boosted regression trees to obtain partial dependence plots. The scoring scheme converted partial dependence plot outputs to the four categories of CVA scoring framework (“low”, “moderate”, “high”, “very high”). Three types of response variable (presence/absence, positive count, all count) were scored to account for various types of monitoring data collected by other monitoring programs. Ages of prickly sculpin were modeled together and separately to discern the importance of accounting for age of the species when performing a sensitivity assessment of monitoring data. Results indicate the data type used for models lead to varying sensitivity scores, especially for seasonal species such as the longfin smelt. Grouping fish by age was important as prickly sculpin showed distinct sensitivity scores for their two age groups. Understanding the ecology and behavior of the fish was also fundamental for understanding model outputs, as demonstrated by results for the threespine stickleback. This study draws attention to the benefits and applicability of long-term monitoring in creating sensitivity analyses for climate vulnerability assessments.