2022 Early Hearing Detection & Intervention Virtual Conference

March 13 - 15, 2022

<< BACK TO AGENDA

5/27/2021 | 2:00 PM - 3:30 PM | Estimating observed-to-expected taxonomic richness in streams and lakes to support existence valuation of biological health | Virtual Platform

Estimating observed-to-expected taxonomic richness in streams and lakes to support existence valuation of biological health

The U.S. EPA is striving to quantify the existence value of biological health in the Nation’s streams and lakes through stated preference (SP) surveys. Through focus groups, we concluded that the public understood observed-to-expected taxonomic richness (O/E) better than other candidate indicators of biological health. To quantify how value varies with proximity to resource, we need spatial interpolations of O/E for all perennial streams and lakes within the conterminous US. To make these interpolations, we used random forest to model O/E with data from the National Aquatic Resource Surveys and StreamCat/LakeCat datasets for the Eastern, Plains, and Western ecoregions. Benthic macroinvertebrates (BMI) and plankton were used for streams and lakes, respectively. The models explained 25-30% of the variation in BMI O/E scores (RMSE: 0.25-0.27) and 13-36% in plankton O/E scores (RMSE: 0.22-0.25). Maps of model residuals showed no visual spatial biases, however, there were noticeable differences in predicted values at ecoregion boundaries. To improve model performances, we are exploring modeling O and E separately, using different modeling algorithms, and including new landscape variables. We will also discuss challenges and solutions for combining model interpolations from streams and lakes.

Landscape
Machine learning
Big data

Presentation:
This presentation has not yet been uploaded.

Handouts:
Handout is not Available

Transcripts:
CART transcripts are NOT YET available, but will be posted shortly after the conference

Presenters/Authors

Jessie Doyle (), ORISE Research Fellow, jdoyle@csumb.edu;