|Valerie Pasquarella||Combining novel remote sensing methods with FIA data to evaluate effects of drought and gypsy moth defoliation on tree mortality at landscape scales||Tuesday||2b||Salon D/E||Integration of Forest Health Data Sources: Leveraging Big Data Tools to Shape the Future of Sharing Information||Gypsy moth defoliation is contributing to declining abundance of oaks in the Northeast. We combine Landsat-based estimates of defoliation severity with FIA plot data to examine relationships between defoliation, drought, mortality, and the net growth of trees. We present a new Google Earth Engine toolset for mapping and monitoring the intensity of gypsy moth defoliation based on new and historic Landsat imagery, as well as an evaluation of regional-scale forest condition assessment products.||Gypsy moth (Lymantria dispar) defoliation is contributing to declining abundance of oaks in Eastern North America. Stand-level observations suggest drought predisposes trees defoliated by gypsy moths to mortality, however the extent to which drought and defoliation have joint effects on tree mortality and growth at regional scales is less understood. We combine Landsat-based estimates of defoliation severity with FIA plot data to examine relationships between defoliation, drought, mortality, and the net growth of trees.
In order to provide consistent, repeatable assessments of changes in forest condition over time, we utilize a new remote sensing approach to map and monitor the intensity of gypsy moth defoliation based on new and historic Landsat imagery. Because this approach is data-intensive, it has previously only been applied to parts of New England. To assess changes in forest condition over larger areas, we have implemented in Google Earth Engine, and this presentation will focus on this new toolset and our evaluation of regional-scale defoliation products. We also highlight challenges in combining plot re-measurements with the Landsat record, as well as preliminary analysis of defoliation and tree mortality.
|Christopher Asaro||Estimating baseline mortality for white pine in the central Appalachians to assess the impact of a novel scale insect/pathogen complex||Tuesday||2b||Salon D/E||Integration of Forest Health Data Sources: Leveraging Big Data Tools to Shape the Future of Sharing Information||In the Central and Southern Appalachians, a novel and poorly understood die-back phenomenon in white pine (Pinus strobus L.) has been attributed to a scale insect and and weak fungal pathogen. Monitoring plots established in 2012 show considerable mortality primarily in younger age cohorts in mixed pine-hardwood stands. After 5+ years of monitoring, baseline mortality analysis indicated that five (2 cm) diameter classes had observed mortality that was significantly greater than baseline.||In the Central and Southern Appalachians, a novel and poorly understood die-back phenomenon in white pine (Pinus strobus L.) has been attributed to a scale insect and and a weak fungal pathogen. Monitoring plots established in 2012 show considerable mortality, primarily in younger age cohorts in mixed pine-hardwood stands. While overall decline in white pine volume has been minimal thus far, baseline mortality analysis indicated that five (2 cm) diameter classes (4-6, 6-8, 8-10, 14-16, 24-26) had observed mortality that was significantly greater that the established baseline mortality after only 5 years of monitoring. Initially, only one of these diameter classes (4-6) had observed mortality significantly greater than the baseline. Baseline mortality across all white pine cohorts was estimated to be approximately 12% based on study plot data and 13-14% based on FIA data. Observed mortality on study plots was over 30% for the smallest diameter classes (4-6, 6-8), and over 20% for the other three diameter classes. If these trends continue, white pine abundance or sustainability in these mixed stands could eventually be compromised as mature trees gradually die and fewer cohorts of younger trees are available to replace them.|
|Erin Berryman||Rating FIA plots for southern pine beetle hazard and applications for the National Insect and Disease Risk Map||Tuesday||2b||Salon D/E||Integration of Forest Health Data Sources: Leveraging Big Data Tools to Shape the Future of Sharing Information||Here, we explore the use of plot data obtained from the Forest Inventory and Analysis database as inputs to southern pine beetle hazard models. Model output is presented and compared to predictions in the most recent National Insect and Disease Risk Map. A possible future outcome of this work is a risk rating system for plots that can be added as a supporting table in FIADB.||The National Insect and Disease Risk Map (NIDRM) estimates areas and magnitudes of tree mortality due to major insect and disease (pest) outbreaks, over a 15 year future time horizon. There is a need for annual assessments of hazard. Many NIDRM models are based on metrics that can be solely derived from FIA plot measurements. We applied southern pine beetle (SPB) models to FIA data. Outputs for each FIA plot location were relative hazard rating (0-10, unitless), and total predicted percent basal area loss over the next 15 years. Because of the importance of geographic representation of model outputs for making adjustments and coordination with state and regional entomologists and pathologists, we compared two scales for spatially-averaging plot hazard: counties, and 6000-acre hexagonal lattices which included the 3-4 nearest plot locations for each hexagon using a buffer of 2600 m outside the edge. We compare results to the most recent (2018) NIDRM model for SPB and explore opportunities for including additional FIA measurements that would allow us to develop more complex models of insect and disease hazard. A possible future outcome of this work is a risk rating system for plots that can be added as a supporting table in FIADB.|
|Kevin Potter||Insect and disease threats to United States tree species and geographic patterns of their potential impacts: Combining FIA data with agent-specific distributions and severities||Tuesday||2b||Salon D/E||We identified the most serious insect and disease threats to each of 419 tree species native to the conterminous 48 United States. By combining the information from this list with FIA data, we were able to assess the potential ecological impacts of important forest insects and diseases within a geospatial context. In general, the potential impacts of insects and diseases were greater in the West. The potential impact of exotic invasive agents was greater in the East, however.||Diseases and insects arguably pose the most destructive threat to North American forests. As part of an effort to identify United States tree species and forests most vulnerable to these epidemics, we compiled a list of the most serious insect and disease threats for 419 native tree species and assigned a severity rating for each of the 1378 combinations between mature tree hosts and 339 distinct insect and disease agents. We then joined this list with Forest Inventory and Analysis data to assess the potential ecological impacts of insect and disease infestations. Specifically, we used the potential host species mortality for each host/agent combination to weight species importance values on approximately 132,000 FIA plots. When summed on each plot, these weighted importance values represent an estimate of the proportion of the plots existing importance value at risk of being lost. In general, the potential impacts of insects and diseases were greater in the West, where there are both fewer agents and less diverse forests. The impact of non-native invasive agents, however, was potentially greater in the East.|
|Susan Crocker||Using forest inventories to assess long-term change in the structure and composition of the Northeastern ash resource in relation to the EAB invasion||Tuesday||2b||Salon D/E||Here, the long-term structural and compositional response of northeastern US ash forests to the invasion of EAB was examined. Using a measure of EAB presence based on county-level detections and tree-ring records, we evaluated changes in ash density, basal area, and regeneration in states that were first affected by EAB (Michigan, Indiana, and Ohio) to determine impact trajectories in recently invaded states. Results are based on analysis of 12,363 FIA plots measured 3x between 2001 and 2017.||While insects are natural disturbance agents that help regulate forest ecosystem processes, the introduction of nonnative insects can significantly alter forest structure and composition, and in turn, stand productivity and biodiversity. Originating in Asia, emerald ash borer (EAB), Agrilus planipennis, Coleoptera: Buprestidae, was introduced to southeastern Michigan in the early to mid-1990s, though it was not detected until 2002. In this study, the long-term structural and compositional response of northeastern United States (US) ash forests to the invasion of EAB was examined. Using a measure of EAB presence based on county-level detections and tree-ring records, we evaluated changes in ash density, basal area, and regeneration in states that were first affected by EAB (Michigan, Indiana, and Ohio) to determine impact trajectories in recently invaded states. Results are based on analysis of 12,363 FIA plots that were measured three times between 2001 and 2017. Quantifying the cumulative changes to ash forests in the northeastern US since the arrival of EAB will help inform land management practices designed to minimize the impacts of this insect.|
|Randall Morin||Carbon losses resulting from insect and diseases invasions in USA forests||Tuesday||2b||Salon D/E||Worldwide, forests are increasingly affected by non-native insects and diseases, many of which cause substantial tree mortality across large regions. Here we analyze forest inventory plots distributed across the USA to estimate the composite impact of the 15 most damaging forest insects and diseases on total tree biomass and carbon loss. From 2009-2015, all non-native forest pest species combined induced an average annual mortality rate (above background mortality) of 4.7% of living tree biomass||Worldwide, forests are increasingly affected by non-native insects and diseases, many of which cause substantial tree mortality across large regions. North America has been invaded by a particularly large number of such species; over 400 forest insect and pathogen species are known to be established and damage has been recorded from about 70 of these. While information exists about ecological impacts of certain individual species, a comprehensive analysis assessing the composite ecosystem impacts of these species is lacking. Here we analyze > 120,000 forest inventory plots distributed across the USA to estimate the composite impact of the 15 most damaging forest insects and diseases on total tree biomass and carbon loss. We determine that the most damaging agents, in terms of biomass and carbon loss, are Dutch elm disease, emerald ash borer, beech bark disease, and hemlock woolly adelgid. From 2009-2015, all non-native forest pest species combined induced an average annual mortality rate (above background mortality) of 4.7% of living tree biomass. These results indicate that forest pest invasions, driven primarily by globalization, are contributing to the accumulation of atmospheric greenhouse gases.|
|Bradley Lalande||Assessment of landscape-level microbial community diversity throughout FIA monitoring plots||Tuesday||2b||Salon D/E||Soil samples were collected during the 2018 field season in conjunction with 49 FIA plots located within Colorado and Wyoming. Soils were sampled in association to all bulk density/duff and litter collections (125 samples). Soil DNA was extracted and sequenced at the 16S (bacterial) and ITS2 (fungal) regions to identify specific microbial communities. Analyses will be prepared to assess the relationship between soil microbes, forest ecotypes, and other relevant site and stand characteristics.||Soil microbes are primary drivers of forest ecosystem processes and play important roles in nutrient cycling, soil structure formation, decomposition, detoxification, regulating plant productivity, influencing plant diversity, and suppression of root diseases. Soil microbial communities are critical to the forest health components of sustainability, productivity, and resilience. However, our understanding of their synergistic interactions tends to be localized and difficult to apply to the landscape or ecoregion level. Soil samples were collected during the 2018 field season in conjunction with 49 FIA plots located within Colorado and Wyoming (125 samples). Using molecular approaches, we can determine taxonomic diversity of microbial communities directly from DNA extracted from environmental soil samples. The identification of microbial communities will increase our understanding of landscape level microbial diversity in forested ecosystems. Additionally, documenting potential drivers of microbial communities (i.e. changes in silvicultural treatments, forest types, land use, and regional climate change) across landscapes and over time is a valuable resource for modeling/predicting and managing western forests.|
|Chad Babcock||Remote sensing-assisted multivariate hierarchical spatial models for simultaneous prediction of multiple forest carbon pools||Wednesday||2b||Salon D/E||Advances in estimating carbon stocks and fluxes on forest land, woodlands, and urban trees across space and through time||We build a multivariate hierarchical spatial model to simultaneously predict six forest carbon pools across the continental US. Leveraging Landsat remote sensing information and cross-correlations between separate carbon pools, we are able to improve spatial prediction and generate small-area forest carbon estimates with better precision than current FIA procedures that use only field plot data.||The FIA monitors six forest carbon pools by taking repeat measurements on plots, including (1) live tree, (2) understory, (3) standing dead, (4) downed dead wood, (5) litter and (6) soil organic carbon. Some pools, such as live tree and standing dead, can be highly correlated with remote sensing data. However, other pools, such as litter and soil organic carbon, can be very weakly correlated with remote sensing products. All six forest carbon pools themselves typically are highly related, i.e., cross-correlated. These cross-correlations are useful information that, if appropriately modeled, can improve remote sensing-assisted prediction for all six pools. We demonstrate how a multivariate hierarchical spatial model can be paired with remote sensing data to predict multiple forest carbon pools simultaneously. Because we are using a unified multivariate framework for spatial prediction, we can explicitly model the cross-correlations between the six carbon pools. Our model is informed by FIA forest carbon pool plot data as response variables and Landsat-derived metrics as explanatory variables. Using this model framework, we make 30 meter spatial predictions for all NLCD-classified forest areas in the continental US.|
|Lucia Fitts||Effects of disturbances and land use change on carbon stocks in six US states||Wednesday||2b||Salon D/E||Advances in estimating carbon stocks and fluxes on forest land, woodlands, and urban trees across space and through time||This research focuses on how disturbances and land use change affect carbon stocks by first doing a model using a machine learning algorithm and a generalized mixed effects model to predict the probability of land use change in FIA plots from 2000 to 2017. Predictions from these models were used to create a regression with spectral information from Landsat imagery and produce a wall-to-wall probability map of land use change.||Forests serve as a large terrestrial carbon sink. Land use change is a major threat to forested areas, but the likelihood of change in forested areas is unknown. This research focuses on forest as a land use, and how forest conditions and population growth affect the carbon stocks in six U.S. states. The goals of the study are (1) to quantify the impacts of forest disturbances on the probability of forest becoming non-forest and (2) to quantify the resiliency of forest carbon stocks when faced with forest disturbances. FIA data was used to generate models with two approaches: a machine learning algorithm (random forest), and generalized mixed effects models. Predictions from these models were used to create a regression with spectral information from Landsat imagery and produce a wall-to-wall probability map of land use change. Results indicate that land use change from forests is more likely with increasing population and housing growth rates, and non-public areas have a higher probability of forest change. Furthermore, disturbances were not a major predictor of land use change. This study provides information for decision makers and land managers for designing policies and practices aimed at mitigating climate change.|
|Richard Birdsey||A User Protocol for Greenhouse Gas Inventories of Forests and Trees at Community Scale: Combining Forest Inventory and Remote Sensing||Wednesday||2b||Salon D/E||Advances in estimating carbon stocks and fluxes on forest land, woodlands, and urban trees across space and through time||A new GHG protocol is designed to support forestry-related decisions by communities for achieving net reductions in emissions. Land management decisions are typically made at small scales; yet, barriers to accessing and using data and models has prevented consideration of land management options. We have designed and tested a protocol that can be used by local planners to access publicly available data and compile estimates of GHG emissions and removals for forests and trees outside forests.||A new greenhouse gas (GHG) protocol is designed to support forestry-related decisions by communities (i.e., municipalities, counties, land-based cooperatives, large ownerships, states) for the purpose of achieving net reductions in GHG emissions. Land management decisions are typically made at relatively small scales such as U.S. counties; yet, barriers to accessing and using data and models has resulted in a lack of consideration of land management options in the hundreds of city and county climate action plans across the country. Working closely with city, county and municipal stakeholders, we have designed and tested a protocol that can be used by local planners to access publicly available data and compile estimates of GHG emissions and removals for forests and trees outside forests. Foundational data include the National Land Cover Data (NLCD) and Forest Inventory Data (FIA), both readily accessible for all areas of the conterminous U.S. The results of three case studies demonstrate how the protocol works. This protocol is associated with ICLEI - Local Governments for Sustainability, whose ClearPath is the online software platform used by US communities to report greenhouse gas inventories.|
|Nancy Harris||Mapping global climate mitigation potential from reforestation||Wednesday||2b||Salon D/E||Research suggests that reforestation is the single largest natural solution for mitigating climate change, but the geographic distribution of reforestations sequestration opportunity remains poorly characterized. Here we combine plot networks with machine learning to map near-term forest carbon sequestration potential with more geographic specificity than the regional estimates that currently exist.||Recent research suggests that reforestation is the single largest natural climate solution for mitigating climate change, and many countries have prioritized forest activities in their Nationally Determined Contributions (NDCs) to the Paris Agreement. However, the magnitude and geographic distribution of reforestations sequestration opportunity remains poorly characterized. One contributing factor is limited data on rates of biomass accumulation, which are spatially highly variable and influenced by numerous biophysical and management factors. By combining existing georeferenced field plot networks from around the globe (including FIA plots), biophysical predictor variables, and machine learning, we mapped current and potential future carbon sequestration potential of natural forest regeneration over the next 20-30 years with more geographic specificity than the regional estimates that currently exist. This type of information can help improve existing GHG inventories at subnational scales where management decisions often take place, and enable local decision makers to evaluate the feasibility, time frame, magnitude, and extent to which reforestation could contribute to and enhance NDCs and local Climate Action Plans.|
|Glenn Christensen||U.S. West Coast Forest Carbon Stocks and Flux: Results from recent analysis and using FIA data as basis informing state-led forest carbon emission mitigation policies||Wednesday||2b||Salon D/E||Following Californias lead, Oregon and Washington along the U.S. West Coast increasingly seek to adopt climate goals through specific strategies reducing carbon (C) emissions. The PNW Research Station developed new partnerships with Oregon and Washington that use FIA data to assess current C stocks and flux on forest land and estimate statewide rates of carbon dioxide (CO2) sequestration. Findings by state including notable differences in growth and mortality in the tree C pool are presented.||Following Californias lead, Oregon and Washington along the U.S. West Coast has increasingly sought to adopt climate goals through specific strategies to reduce carbon (C) emissions. Understanding the role forests play in each states overall C stocks and flux is critical to developing, implementing, and monitoring these strategies. Similar to the role FIA continues to play in California and our partnership with the California Department of Forestry and Fire Protection (CAL FIRE), the Pacific Northwest Research Station has developed new partnerships with the Oregon Department of Forestry and Washington State Department of Natural Resources to assess current C stocks and flux on forest land in order to estimate a statewide rate of carbon dioxide (CO2) sequestration. We present the latest findings from California, Oregon, and Washington including notable differences in growth and mortality in the tree C pool. The future direction of this work are discussed including state-led efforts improving the reporting and confidence of FIAs forest C stock and flux estimates.|
|Todd Morgan||Leveraging TPO data for HWP carbon storage in the PNW||Wednesday||2b||Salon D/E||Stakeholders in FIAs Pacific Northwest (PNW) region have interest in quantifying carbon (C) stored in harvested wood products (HWP) because of state climate goals regarding C emissions. University of Montana researchers have used timber product output (TPO) information and an IPCC based carbon storage model developed by the Forest Service to quantify C stored in landfills and products made from historic and recently harvested timber in California and Oregon, with a Washington study pending.||Stakeholders in FIAs Pacific Northwest (PNW) region have interest in quantifying carbon (C) stored in harvested wood products (HWP) because of state climate goals regarding carbon emissions. The FIAs Timber Product Output (TPO) program quantifies and characterizes timber used for products and disposition of mill residue, using mill surveys and logging utilization studies. In the PNW Region, this survey work is conducted by the University of Montanas Bureau of Business and Economic Research (BBER). Working with PNW-FIA, BBER researchers are using the Intergovernmental Panel on Climate Change (IPCC) production accounting approach to estimate HWP-C storage in California, Oregon, and Washington with a model developed by the Forest Service as well as timber harvest, primary products and residue generated by timber-processing facilities. Results from California and Oregon indicate HWP-C storage is a fraction of ecosystem C storage. However, being able to consistently quantify C stored in HWP through time and across political boundaries will allow stakeholders and policy makers to better understand the role that timber harvested and used for wood products plays in total forest carbon storage and flux.|
|Jeremy Fried||Propensity score matching analysis of FIA remeasurement across diverse West Coast forests yields insights on drivers of in-forest carbon sequestration||Wednesday||2b||Salon D/E||Marginal impact of stand structure, management and disturbance on above-ground net carbon increment was analyzed via propensity score matching on NFS and private FIA plot subsets in WA, OR and CA. Carbon increment associated positively with active management, site quality, and initial inventory everywhere. Even-aged stand structure elevated increment in coastal PNW. Fire, insect, and disease reduced increment inland. High QMD reduced stand level carbon accumulation everywhere.||To estimate stand structure, management and disturbance influence on above-ground net carbon increment on timberland, we developed matched subsets totaling 3920, not recently harvested, remeasured, forest service and private FIA plots in WA, OR and CA. Grouping plots into four regions (PNW coast/inland and PSW coast/inland), we applied propensity score matching (PSM) to generate comparable plot subsets for a better apples to apples evaluation of the marginal influence of those drivers on above-ground live tree carbon increment. As expected, carbon increment associated positively with site quality and initial inventory everywhere. Even-aged stand structure associated with higher increment only in PNW coastal. Negative impacts of fire, insect, and disease were evident in only inland regions. Higher quadratic mean diameter of trees reduced carbon increment in all regions. The more active forest management common on private timberlands elevated carbon increment in all regions, all else equal. PSM presents a promising approach for creating datasets amenable to evaluating the effect of discrete variables that arent randomly distributed across the forested landscape but are key to testing policy relevant forest management hypotheses.|