Lauren N. Youngblood | Geology Reports

USGS2026

Estimating basal area change by tree size with Sentinel-2 imagery following four fires in California, USA

Background Failure to account for tree size when estimating burn severity may not accurately capture post-fire tree mortality and post-fire forest structure. Aims We explored whether basal area mortality by tree size class could be determined from remotely-sensed burn severity indices based solely on Sentinel-2 satellite imagery. Methods We used data collected in four large California wildfires to model the relationship between proportional basal area mortality and burn severity indices derived from Sentinel-2 imagery for three tree diameter class thresholds: small (15 to 30 cm), medium (30 to 50 cm) and large (>50 cm). Key results Our models showed that for a given burn severity index value, the proportion of mortality was greater overall in smaller trees, and that the proportion of mortality in large trees changed more slowly than that of smaller trees with changing burn severity index values. Conclusions We found that models that accounted for tree size can more precisely estimate changes in forest size structure than a similar model that did not account for tree size. Implications Explicitly accounting for tree size can improve estimates of post-fire forest structure, including for large trees which make up the bulk of stand biomass and post-fire seed sources.

California↗

USGS2025

Vegetation community monitoring: Forest structure in Klamath Network parks

The Klamath Network, comprising six national park units in northern California and southern Oregon, initiated a vegetation monitoring protocol in 2011 to discern ecologically significant vegetation trends in these parks. The premise of the protocol is that multivariate analyses of species composition data is the most robust means for early detection of vegetation change over time. Here, we present these community metrics based on our sampling efforts from 2011 to 2019. Observations from the first sampling event (2011–2013) were used to establish baseline conditions for the vegetation communities. Observations from subsequent sampling in 2014–2019 were used to identify potential temporal variation in forest structure across habitat types and parks. Park landscapes were categorized into three strata: matrix (low- to mid-elevation upland habitats), riparian (within 10 m of a perennial stream), and high-elevation (above a predefined elevation, park-specific). At the onset of the network’s vegetation monitoring protocol, 241 permanent plots were established at random locations across the three strata. We present summary statistics from three repeated samplings (2011–2019) of each plot, describing variation in forest structure across broad habitat types and parks. Observable differences in forest structure aligned with expected productivity gradients across the parks. Measures of forest structure (vegetation cover, stem density, basal area, tree heights, height to live crown, shrub cover, and surface fuels) were generally higher in mesic sites, compared to sites located in more arid, continental climates. Differences across sampling frames also followed this general pattern of productivity. Matrix and riparian sampling frames had similar ranges of values in most cases, while high elevation sites had relatively lower stem density, basal area, shrub cover, fuels, and recruitment. Notably, we observed a relative lack of change in forest structure over time. This is not surprising given the relatively short (six-year) timespan of observations in each park. The fourth set of Klamath Network surveys (2021–2023) is likely to show substantial changes in vegetation cover and forest structure, particularly for parks that have recently experienced major fires. Continued long-term vegetation monitoring is crucial for understanding ecosystem responses to a rapidly changing world. This report on vegetation composition is the second in a series; upcoming reports will analyze structure and function, aiming to detect spatiotemporal trends.

California, Oregon↗