Vegetation Mapping and Monitoring 2

Friday, October 21 at 10:00-11:40 am, Fir Room

*note alternate instance of this session – Friday at 8am

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

17.1 A general framework for identifying the impact of vegetation disturbance using Landsat Timeseries in Santa Barbara County, California

Dar Alexander Roberts (University of California, Santa Barbara, Santa Barbara, CA, USA), Christopher Kibler (University of California, Santa Barbara, Santa Barbara, CA, USA), Conor McMahon (University of California, Santa Barbara, Santa Barbara, CA, USA)

Widely available, high quality, standardized remote sensing products have become increasingly important for identifying surface changes due to disturbance over broad spatial and temporal scales. Spectral Mixture Analysis (SMA) decomposes the mixed reflected signal from multiple materials in a pixel into estimates of fractional cover of pure components, called endmembers. For forests and shrublands these are typically Green Vegetation (GV), Non-photosynthetic Vegetation (NPV, stems, litter, wood), Soil, and Shade. When applied to time series, SMA not only identifies that an event has occurred, but also provides a physically-based measure of change, such as GV loss and increased NPV due to fire or forest pathogens. We present a generalized framework for applying SMA to annual time series to calculate the pixelwise mean and variance of each cover fraction and define average cover over time. Disturbance events are identified by subtracting the long-term mean from cover fractions for individual years, generating a residual image that defines the type of cover change, the direction of change and its magnitude. We illustrate the potential of this approach using Landsat data acquired over Santa Barbara County, California, from 1984 to 2021, using anniversary dates primarily from July. We find inter-annual variation in precipitation and fire to be the dominant sources of change across the landscape; high rainfall years generated an increase in GV, while drought produced decreased GV and increased NPV due to canopy dieback. Upland areas were more sensitive to drought than riparian areas. Extensive positive NPV anomalies were observed throughout the extreme 5-year drought of 2012–2016, suggesting a build-up of dry fuels. Major fires generated the largest changes, resulting in decreased GV and increased NPV fractions that persisted for five to seven years following fires. Following the Thomas Fire of 2017, the 2018 Montecito Debris Flows were expressed as a positive Soil and NPV residual in impacted watersheds.

17.2 A new era for Calveg mapping: accelerating inventory and technology for map updates

Michèle Slaton (USDA Forest Service Region 5, Remote Sensing Lab, McClellan, CA, USA)

The CALVEG existing vegetation map product serves as a foundation for land management planning and monitoring across multiple ownerships. The current refresh cycle of 10-20 years, overlaid with recent broad-scale forest mortality events and megafires has resulted in sorely outdated map products. To fill this gap, the Forest Service has developed a field plot and Landsat-based imputation system known as “F3” (Forest Inventory & Analysis plots, Forest Vegetation Simulator, and Fastemap) to update existing vegetation map products. We first report on a pilot project on the San Bernardino National Forest and in the Sierra Nevada using this novel method. Secondly, we report on the utilization of a LiDAR collection to inform forest structural attributes in the same project areas, which may also be expanded to other national forests after broad-scale LiDAR collections are conducted this year and next. Our findings demonstrate greater efficiency of mapping through automation, especially in a rapidly changing environment, but also raise important questions about needed revisions to the CALVEG classification. Plans for existing vegetation updates on other national forests are outlined, as are the challenges of conducting restoration project planning while map updates remain underway.

17.3 Declining southern Sierra Nevada conifer forests in an age of mega-disturbances: applying remote sensing tools to inform mature forest conservation

Zack Steel (University of California, Berkeley, Berkeley, CA, USA), Gavin Jones (USDA Forest Service, RMRS, Albuquerque, NM, USA), Brandon Collins (University of California, Berkeley, Berkeley, CA, USA), Rebecca Green (USDA Forest Service PSW, Davis, CA, USA), Alexander Koltunov (USDA Forest Service Region 5, McClellan, CA, USA), Kathryn Purcell (USDA Forest Service PSW, Coarsegold, CA, USA), Sarah Sawyer (USDA Forest Service Region 5, Vallejo, CA, USA), Michèle Slaton (USDA Forest Service Region 5, McClellan, CA, USA), Scott Stephens (University of California, Berkeley, Berkeley, CA, USA), Peter Stine (Stine Wildland Resources Science, Sacramento, CA, USA), and Craig Thompson (USDA Forest Service PSW, Fresno, CA, USA)

Mature forests – characterized by high cover of tall trees and complex understories – are important habitats for native plant and wildlife species and support critical ecosystem functions globally. In California’s Sierra Nevada, a combination of a century of fire exclusion and worsening climate change has led to increasingly severe wildfires, and ongoing extreme drought threatens habitats of sensitive species. Using spatially explicit datasets of forest structure (the “F3” data product) and the Ecosystem Disturbance and Recovery Tracker (eDaRT), we quantified the loss of conifer forest cover in the southern Sierra Nevada between 2011 and 2020, a region and decade characterized by unprecedented mega-disturbances. Due to the combination of wildfires, drought, and drought-associated mortality from bark beetles, 30% of conifer forest area transitioned to non-forest vegetation (i.e., canopy cover fell below 25%) during this period. Of the spatially limited mature forest habitats, 56% of moderate density (40-60% canopy cover) and 84% of high density (>60% canopy cover) forests transitioned to lower density forests or non-forests. Drought and beetle-kill caused greater cumulative declines than areas where wildfire mortality overlapped with the other disturbances. However, burned areas resulted in larger patches of forest loss and greater forest fragmentation on average. These results highlight that current conservation and management approaches are failing to protect mature forest habitats within disturbance-prone ecosystems like the conifer forests of California. We emphasize the need to switch from a static approach toward one focused on managing healthy disturbance dynamics, especially using frequent low-severity fire to increase resilience of forests to future mega-disturbances. Without rapid management interventions, remaining mature forest habitat may be susceptible to complete loss in the coming decades.

17.4 Geographic approach for Natural Areas management

Sunny Fleming (Esri, Redlands, CA, USA)

Natural Areas provide vital protection for species and the habitats they call home. They perform ecosystem services and contribute to our own quality of life through recreational opportunities. Managing these areas in a changing climate has become increasingly challenging as threats such as natural disasters and invasive species are exacerbated. Ensuring we can stay ahead of these challenges requires innovating how we use technology to conduct the work of managing these lands. In this presentation, we look at examples and ideas for how Esri’s geographic information systems (GIS) technology improves our operations in the field and in the office, help us prioritize and predict, and lead to more informed decisions that help us achieve our goals of a more sustainable future.