ATTACHMENT_02_-_140P8623Q0107_-_SOW.pdf

REQUEST FOR QUOTATION (RFQ) 140P8623Q0107

SCOPE OF WORK (SOW)

Title: Analyses of Vegetation Trends and Statistical Power for MEDN I&M

Project Description:

The Mediterranean Coast Network parks (Channel Islands National Park, Santa Monica

Mountains National Recreation Area, and Cabrillo National Monument) systematically monitor vegetation according to the MEDN I&M terrestrial vegetation monitoring protocol

(Tiszler et al. 2016).

According to the monitoring protocol (Tiszler et al. 2016), an in-depth synthesis report on status and trends will be produced at the end of every monitoring cycle. Reports for CABR and CHIS have been completed, and we only need the trend report for SAMO. Up to 200 outcomes of interest are available and include species-specific outcomes; vegetation alliances and functional groups such as trees, shrubs, forbs, and grasses; and community metrics such as species richness and diversity. Covariate information such as precipitation, temperature, soils, slope aspect, wind direction, and solar radiation will be considered as explanatory variables in trend models.

The first task involves conducting a trend analysis for a set of 12 outcomes of interest (to be determined at the project kickoff meeting) collected at SAMO through 2022. This analysis will model trend as a function of time and as a function of covariates in separate models. The second task involves conducting power analyses for 12 outcomes of interest at each park using simulated data and baseline status and variance components estimates from the recently completed trend analyses. A single set of variance components for each outcome of interest using the results from the trend analyses will be used for both the index sites and the probabilistic sites at CHIS. The power analysis will include a small set of simulation scenarios to assess the performance of the current sample size using the current temporal revisit designs at each park.

Sampling Background:

The Tiszler et al. (2016a, 2016b) vegetation monitoring protocol was a newly implemented program at SAMO and CABR, but primarily a change in site selection approach at CHIS, which began in 1984. Monitoring sites on Anacapa Island (ANI), Santa Barbara Island (SBI), San

Miguel Island (SMI), and Santa Rosa Island (SRI) were subjectively selected “index” (of change) sites within major plant communities, using relevé data and vegetation maps for the purpose of examining recovery after nearly a century of grazing. Sites on Santa Cruz Island

(SCI) were selected using stratified random sampling, with stratification based on watersheds

(Tiszler et al 2016a, Appendix A). While the index sites are important for monitoring and understanding vegetation dynamics on SBI, ANI, SRI, and SMI, they do not allow for statistical inference to all vegetation on the islands. Under the newer protocol, CHIS staff continue monitoring the 181 established index sites, although on a less frequent basis, and added 184 new probabilistically selected monitoring sites that will provide statistical inference to all vegetation at CHIS.

At SAMO, 100 sites (two panels of 50 each) are monitored each year for a total of 300 sites in a

6-year cycle (6 panels). At CABR, 24 sites (two panels of 12 each) are monitored each year for a total of 72 sites.

The new monitoring sites at CHIS and all of the monitoring sites at SAMO and CABR were selected using Generalized Random Tessellation Stratified (GRTS) sampling with equal probability weighting (Stevens and Olsen 2003, 2004). The CHIS revisit design uses a [1-0, 1-3] augmented serially alternating panel design. (Notation is from McDonald (2003): the first number indicates the number of years a site is monitored before resting and the second number indicates the number of years a site is rested before resuming monitoring). Each panel will include existing index sites and new probabilistically selected sites in approximately equal numbers. The SAMO and CABR revisit designs follow a [2-4] serially alternating panel design.

Each of six panels is monitored on a 6-year cycle in which all sites in the panel are visited for two consecutive years and then rested for four years. The cycles are staggered such that successive panels overlap by one year.

The unbiased estimation of status and trend parameters will be obtained with a variety of design-based, model-assisted, and model-based approaches. Design-based estimation will be based on

GRTS sampling theory, and analysis will employ tools available in the spsurvey R package.

Nonresponse should be addressed by examining the underlying missingness mechanism and choosing the appropriate analysis approach. Trend analysis will incorporate the linear mixed model, but nonparametric approaches may also be applied to provide corroborating evidence of significant trend. When outcomes are zero-inflated, non-linear maximization approaches can be used with zero-inflated models to obtain estimates of trend for both the mean and the proportion of zeros observed. The methods outlined in SOP 14 of Tiszler et al. (2016b) represent a few possible approaches for status estimation and trend testing, but other approaches may be appropriate for the monitoring data and should be considered and applied by qualified personnel so that appropriate inference is made.

Project Objectives:

Analysis of trends in vegetation at SAMO are needed to understand the condition of vegetation and the effects of wildfire and drought. Power analyses are needed for each park to determine if the sampling design is achieving protocol objectives, and if the sampling can be reduced. The objectives of this project are to:

(1) describe trends in vegetation at SAMO

(2) evaluate the effects of potential explanatory value of covariates

(3) synthesize this information to explore patterns and better understand processes of vegetation dynamics and potential drivers of change

(4) evaluate the current sampling design at each park

(5) make recommendations to improve program efficiency and effectiveness

Results will be provided in the form of multiple reports (SAMO Veg trends, and then one power analysis report for each park), in addition to conference calls and webinars with scientists and resource managers. NPS staff will summarize the results for sharing with the public.

Description of Data:

Data will be provided as three datasets, one for each park. All data will be provided in a format that is ready for analysis. Count data in each dataset will have undergone a variety of quality control checks. Metadata explaining the methods used to develop the dataset, a description of each field in the dataset, and any other relevant information needed to interpret the quality of the data will be submitted with the data.

Relevant Documents

MEDN native plant monitoring program documents are located here:

https://science.nature.nps.gov/im/units/medn/publications.cfm?tab=3&ProtocolNativeCommunit ies=open#ProtocolNativeCommunities

TASK 1: Analysis of vegetation trends at CHIS

SAMO trend analysis of 12 outcomes of interest.

Estimated timing: October 2023 – September 2024.

Deliverable: A report containing the R code used in the analysis and an explanation of the background, methods, results, discussion, and recommendations to improve the trend estimates.

Additionally, the results will be presented via a webinar to NPS staff.

TASK 2: Analysis of power to detect trends in vegetation trends SAMO, CABR, and CHIS

A power analysis to determine appropriate sample size of plots for 12 outcomes of interest for each park (including one for each island at CHIS).

Estimated timing: November 2024 – September 2025.

Deliverable: A report containing the R code used in the analysis and an explanation of the background, methods, results, discussion, and recommendations to improve the program’s efficiency and effectiveness. Additionally, the results will be presented via a webinar to NPS staff.

Literature Cited

McDonald, T. L. 2003. Review of environmental monitoring methods: survey designs.

Environmental Monitoring and Assessment 85:277-292.Stevens Jr., D. L., and A. R. Olsen.

2003. Variance estimation for spatially balanced samples of environmental resources.

Environmetrics 14:593-610.

Stevens, D. L., Jr., and A. R. Olsen. 2004. Spatially-balanced sampling of natural resources.

Journal of American Statistical Association 99(465):262-278

Tiszler, J., D. Rodriguez, K. Lombardo, T. Sagar, L. Aguilar, L. Lee, T. Handley, K. McEachern, L. Starcevich, M. Witter, T. Philippi, and S. Ostermann-Kelm. 2016a. Terrestrial vegetation monitoring protocol for the Mediterranean Coast Network—Cabrillo National Monument, Channel Islands National Park, and Santa Monica Mountains National Recreation Area:

Narrative, Version 1.0. Natural Resource Report NPS/MEDN/NRR—2016/1296. National Park

Service, Fort Collins, Colorado.

Tiszler, J., D. Rodriguez, K. Lombardo, T. Sagar, L. Aguilar, L. Lee, T. Handley, K. McEachern, L. Starcevich, M. Witter, T. Philippi, and S. Ostermann-Kelm. 2016b. Terrestrial vegetation monitoring protocol for the Mediterranean Coast Network—Cabrillo National Monument, Channel Islands National Park, and Santa Monica Mountains National Recreation Area: Standard

Operating Procedures, Version 1.0. Natural Resource Report NPS/MEDN/NRR—2016/1296.1.

National Park Service, Fort Collins, Colorado.