Vegetation Mapping and Monitoring 1

Friday, October 21 at 8:00-9:40 am, Fir Room

*note alternate instance of this session – Friday at 10am

Session Description: Government agencies, NGOs, academic institutions, and consulting firms continue to improve standards, techniques, and resulting products of vegetation mapping — especially since Geographic Information Systems, imagery, LiDAR, and remote sensing technologies have expanded from the late 20th Century on through today. Vegetation mapping and monitoring are important tools for species, habitat, and landscape-level assessment, analysis, monitoring, and conservation, driving many of today’s decisions for land-use planning. This session showcases promising uses of vegetation mapping and monitoring to positively impact decision-making in conservation and management throughout California.

Session Chairs: Julie Evens (California Native Plant Society, Sacramento, CA, USA) and Michèle Slaton (USDA Forest Service Pacific Southwest Region, Bishop, CA, USA)

*Session generously sponsored by Esri

12.1 Spatial phylogenetics: an important tool for conservation-driven land management in California

Dr. Israel Borokini (University and Jepson Herbaria, University of California, Berkeley, Berkeley, CA, USA), Brent Mishler (University and Jepson Herbaria, University of California, Berkeley, Berkeley, CA, USA)

Habitat loss remains the greatest threat to biodiversity at all geographical scales, which necessitates an emphasis on conservation in land management. Decisions on habitat protection are currently based mainly on species richness and endemism, in addition to the severity of biodiversity threats, cost, and socio-environmental factors. We highlight two limitations of basing conservation prioritization on species richness and endemism. First, “species” represent only one of the many nested levels of biodiversity in the tree of life, thus species-based conservation policies lose other equally-important elements of biodiversity. Secondly, species distribution patterns are non-random products of ecological processes, evolutionary forces, and geological/climatic histories, which are more reliably captured through phylogenetic measurements. Phylogenetic metrics have been developed to describe the patterns of diversity and endemism on the landscape based on the evolutionary uniqueness of the lineages represented in a location, an approach called spatial phylogenetics. This approach combines massive spatial datasets and large phylogenetic trees built with genomic data representing the taxa presented in a study region. These datasets are used to put the phylogeny on the map as a GIS layer and estimate patterns of phylogenetic diversity (PD) and phylogenetic endemism (PE), in relation to other spatial data. Hypothesis tests and tests of statistical significance are made using spatial randomizations. Patterns of PD, PE, and related metrics can be integrated with factors such as land use, intactness, and protection status, using a stepwise algorithm to assess complementarity (that is, how much additional biodiversity would be protected by increasing conservation of a particular area). We demonstrate this method using the California vascular flora and highlight its application to current conservation efforts.

12.2 Remote sensing of post-fire vegetation recovery during extreme drought

Christopher Kibler (University of California, Santa Barbara, Santa Barbara, CA, USA), Anne-Marie Parkinson (University of California, Santa Barbara, Santa Barbara, CA, USA), Dar Alexander Roberts (University of California, Santa Barbara, Santa Barbara, CA, USA), Carla D’Antonio (University of California, Santa Barbara, Santa Barbara, CA, USA), Seth Peterson (University of California, Santa Barbara, Santa Barbara, CA, USA)

Remote sensing and vegetation field surveys provide complementary data sets for monitoring vegetation recovery after wildfires. Remote sensing leverages aerial and satellite imagery to cover broad spatial and temporal extents, however the spatial resolution of the imagery can be relatively coarse and the data are limited to ecological traits that can be reliably related to reflected light measured by the sensor. Vegetation field surveys can provide precise measurements of species cover and plant traits; however, field surveys have limited spatial and temporal coverage. In this study, we combined remote sensing and field surveys to monitor vegetation recovery after the 2007 Zaca Fire in Santa Barbara County, California. Recovery after the fire was affected by the extreme 2012–2019 California drought. Field surveys were conducted in the summer of 2018. We measured species cover in 82 belt transects (of 45 m length) located throughout the burn scar. These data were used to identify the dominant species within each transect. We also analyzed 18 years of Landsat remote sensing imagery (2000–2018) to monitor land cover before and after the Zaca Fire. Spectral mixture analysis (SMA) was used to estimate the fractional cover of green leaves, dead/woody plant material, and soil within each pixel. Our analysis showed a large increase in dead plant material and exposed soil after the fire, which was slowly replaced by green vegetation cover over time. By overlaying the field survey data on the remote sensing data, we identified post-fire recovery trends for different chaparral species. The differences in recovery trends between the species were most pronounced during the 2012–2019 California drought. Species such as Ceanothus cuneatus were highly sensitive to precipitation variability during the drought, while other species such as Cercocarpus betuloides were much less sensitive to drought conditions. Across all species, the drought appears to be substantially delaying the return to pre-fire conditions.

12.3 Mapping native plant species at Abalone Cove

Michael Pazmino (California State University Long Beach MSGISci Program, Long Beach, CA, USA), Minerva Lara (California State University Long Beach MSGISci Program, Long Beach, CA, USA), Joshua Moreno (California State University Long Beach MSGISci Program, Long Beach, CA, USA), Sebastian Alvarez Espinoza (California State University Long Beach MSGISci Program, Long Beach, CA, USA)

First formed in 1988, the Palos Verdes Peninsula Land Conservancy (PVPLC) has preserved over 1,600 acres of undeveloped natural lands for historical, ecological, educational, and recreational purposes. The PVPLC has made it their primary mission to protect at-risk habitats and species including the rare and beautiful Palos Verdes Blue Butterfly, which researchers considered extinct in the early 1980’s before finding again in 1994, along with other endangered species. Over the last several decades, urbanization, invasive plants, and wildfires have expanded extensively in the region. In 2012, the last major fire in Palos Verdes burned over 35 acres of land. Before that, in 2009, 75 acres were burned, and six homes were destroyed. The Los Angeles Fire Department (LAFD) Brush Clearance Unit conducts yearly brush removal operations on properties located in the Very High Fire Hazard Severity Zone (VHFHSZ) to prevent future disasters and wildfires. All lands managed by the PVPLC are in the VHFHSZ, which are required to have yearly brush removal operations. We were tasked to map the location of native vegetation located at the Abalone Cove nature preserve to help guide the PVPLC and local fire agencies on appropriate brush fire prevention operations. We collected Unmanned Aerial Vehicle (UAV) imagery, in January 2022, and we used object-based-identification-analysis to identify the native plant species within our study area and classify their type and species. We used the results with the highest accuracy and created a web map that was shared with PVPLC and others. This map will provide a more comprehensive understanding of the land to guide future habitat restoration and brush removal operations. Upon sharing our methodologies, data, and resulting analyses with CSULB and PVPLC, we will continue to support future projects to map the rest of the PVPLC nature preserves.

12.4 Reference vegetation for ecological restoration and mitigation in Santa Clara County

Claire Mallen (Santa Clara Valley Water District, San Jose, CA, USA), Zooey Diggory (Santa Clara Valley Water District, San Jose, CA, USA)

Reference sites provide a self-sustaining model for the restoration of an ecosystem. Multiple reference sites, which incorporate an ecosystem’s variability in a region, provide a particularly robust basis for restoration goal setting, design, and performance evaluation. This is particularly true for mitigation, which must attain permit-required success criteria. If the project design and success criteria reflect an appropriate range of native ecosystem conditions, then attainment of those criteria should indicate proper ecological function. The use of multiple reference sites is limited, however, by a single project’s or single organization’s ability to survey multiple sites over a reasonable scale in time. In 2020–2021, we documented the composition and conditions of Santa Clara County’s most pristine native plant communities to help make reference site information readily available. At over 300 sample plots a modified Vegetation Rapid Assessment method and Manual of California Vegetation classification system were used to record plant species, percent cover, and key physical conditions. The plots represent multiple sites of 50 different vegetation alliances, ranging from grassland, chaparral and scrub, oak woodland and savanna, conifer woodland, freshwater and tidal marsh, and riparian forest and scrub types. Data can be sorted by location, elevation, soil, and exposure to help identify reference conditions that are most appropriate for a project site. The list of species can inform planting palettes that increase diversity and resilience to climate change; tree and shrub density, relative elevation, and distance from creeks can inform planting plans; and species and growth-form percent cover can inform success criteria. This data, now publicly available, will help improve restoration designs, facilitate ecologically-meaningful permit compliance, and contribute to the resilience and preservation of regional native ecosystems.

12.5 Fine-scale vegetation mapping in the San Francisco Bay Area

Danny Franco (Golden Gate National Parks Conservancy, San Francisco, CA, USA), Mark Tukman (Tukman Geospatial, Forestville, CA, USA), Kass Green (Kass Green and Associates Berkeley, CA, USA)

Over the last eight years, eight San Francisco Bay Area counties have built or are in the process of building fine-scale maps of vegetation, wildland fuels, wildfire risk and other landscape attributes from a combination of high and moderate resolution imagery, field work, and lidar data. Mapping methods use a semi-automated approach that combines machine learning and automation with field validation and manual aerial photo interpretation. With minimum mapping units of 1/4 to 1 acre, the data sets are spatially rich and support multiple applications including wildfire fuels and hazard management, evacuation route planning, disaster response, carbon monitoring, infrastructure planning, watershed management, engineering design, sea level rise adaptation planning, wildlife habitat management, forest management, flood planning and mitigation, and environmental and conservation planning. This talk will provide (1) an overview of the Bay Area wide projects and the grass roots efforts to bring them to fruition, (2) reviews of remote sensing methods developed to create the data sets including field data collection, ancillary dataset development, and machine learning, and (3) examples of how Bay Area agencies, private landowners, and NGO’s are using the data sets to support local and regional decision making and resource management. Fine scale vegetation mapping has been completed for more than 650,000 hectares, with an additional 450,000 hectares to be completed in Summer 2023. Landcover and lidar derived topographic datasets have been accessed and downloaded thousands of times from public data portals by a broad cross-section of end users.