Search

Contemporary forest management requires highly-detailed, spatially-contiguous, multi-temporal, and scenarios-comparable forest conditions. Field inventories and individual-tree models often contain highly-detailed data and allow for long-term complex scenarios comparison, but the information is only at sampled locations and lacks complete spatial coverage. Forest landscape models (FLMs) provide landscape-level spatiotemporal data, but the details that are important to land managers are often lost in the generalized outputs. We developed a modeling framework, F 3 , to integrate FIA (Forest Inventory and Analysis) plots, the Forest Vegetation Simulator (FVS), and FastEmap (Field And SatelliTe for Ecosystem MAPping) to simulate spatiotemporal forest change under natural succession and vegetation management. F 3 extrapolates the details of forest inventory plots and individual-tree model outputs to a spatially-contiguous landscape by fusing tree-list field measurements, individual tree growth and yield models, remote sensing and environmental geospatial datasets. F 3 allows for area-specific management action simulations. F 3 compares FVS results with field measurements for temporal accuracy assessment and uses a leave-one-out cross-validation for spatial accuracy assessment. F 3 adopts parallel computation techniques to implement the modeling in an automatic and efficient manner. The proof of concept of F 3 was demonstrated in Tahoe National Forest (TNF) showing spatiotemporal changes on six forest structural metrics (quadratic mean diameter, basal area, biomass, habitat suitability index, canopy cover, and coarse woody debris) under natural succession, regeneration-cut, and thinning scenarios for the years 2014–2114 at a 30 m resolution. F 3 can be used for initializing FLMs and for analyzing a wide range of ecosystem services; however, the under-representation of certain forest types in the FIA plot data set, the modeling bias from FVS, and choice of FastEmap covariates contribute to major uncertainties in the framework. [ABSTRACT FROM AUTHOR]

High-resolution site index (SI) and mean annual increment (MAI) maps are desired for local forest management. We integrated field inventory, Landsat, and ecological variables to produce 30 m SI and MAI maps for the Tahoe National Forest (TNF) where different tree species coexist. We converted species-specific SI using adjustment factors. Then, the SI map was produced by ( i) intensifying plots to expand the training sets to more climatic, topographic, soil, and forest reflective classes, ( ii) using results from a stepwise regression to enable a weighted imputation that minimized the effects of outlier plots within classes, and ( iii) local interpolation and strata median filling to assign values to pixels without direct imputations. The SI (reference age is 50 years) map had an R2 of 0.7637, a root-mean-square error (RMSE) of 3.60, and a mean absolute error (MAE) of 3.07 m. The MAI map was similarly produced with an R2 of 0.6882, an RMSE of 1.73, and a MAE of 1.20 m3·ha−1·year−1. Spatial patterns and trends of SI and MAI were analyzed to be related to elevation, aspect, slope, soil productivity, and forest type. The 30 m SI and MAI maps can be used to support decisions on fire, plantation, biodiversity, and carbon. [ABSTRACT FROM AUTHOR]

Abstract: In interior Alaska, wildfires change gross primary production (GPP) after the initial disturbance. The impact of fires on GPP is spatially heterogeneous, which is difficult to evaluate by limited point-based comparisons or is insufficient to assess by satellite vegetation index. The direct prefire and postfire comparison is widely used, but the recovery identification may become biased due to interannual climate variability. The objective of this study is to propose a method to quantify the spatially explicit GPP change caused by fires and succession. We collected three Landsat images acquired on 13 July 2004, 5 August 2004, and 6 September 2004 to examine the GPP recovery of burned area from 1987 to 2004. A prefire Landsat image acquired in 1986 was used to reconstruct satellite images assuming that the fires of 1987–2004 had not occurred. We used a light-use efficiency model to estimate the GPP. This model was driven by maximum light-use efficiency (Emax) and fraction of photosynthetically active radiation absorbed by vegetation (FPAR). We applied this model to two scenarios (i.e., an actual postfire scenario and an assuming-no-fire scenario), where the changes in Emax and FPAR were taken into account. The changes in Emax were represented by the change in land cover of evergreen needleleaf forest, deciduous broadleaf forest, and shrub/grass mixed, whose Emax was determined from three fire chronosequence flux towers as 1.1556, 1.3336, and 0.5098gC/MJ PAR. The changes in FPAR were inferred from NDVI change between the actual postfire NDVI and the reconstructed NDVI. After GPP quantification for July, August, and September 2004, we calculated the difference between the two scenarios in absolute and percent GPP changes. Our results showed rapid recovery of GPP post-fire with a 24% recovery immediately after burning and 43% one year later. For the fire scars with an age range of 2–17years, the recovery rate ranged from 54% to 95%. In addition to the averaging, our approach further revealed the spatial heterogeneity of fire impact on GPP, allowing one to examine the spatially explicit GPP change caused by fires. [Copyright &y& Elsevier]

Abstract: Spatiotemporal variations of wetland water in the Prairie Pothole Region are controlled by many factors; two of them are temperature and precipitation that form the basis of the Palmer Drought Severity Index (PDSI). Taking the 196km2 Cottonwood Lake area in North Dakota as our pilot study site, we integrated PDSI, Landsat images, and aerial photography records to simulate monthly water surface. First, we developed a new Wetland Water Area Index (WWAI) from PDSI to predict water surface area. Second, we developed a water allocation model to simulate the spatial distribution of water bodies at a resolution of 30m. Third, we used an additional procedure to model the small wetlands (less than 0.8ha) that could not be detected by Landsat. Our results showed that i) WWAI was highly correlated with water area with an R2 of 0.90, resulting in a simple regression prediction of monthly water area to capture the intra- and inter-annual water change from 1910 to 2009; ii) the spatial distribution of water bodies modeled from our approach agreed well with the water locations visually identified from the aerial photography records; and iii) the R2 between our modeled water bodies (including both large and small wetlands) and those from aerial photography records could be up to 0.83 with a mean average error of 0.64km2 within the study area where the modeled wetland water areas ranged from about 2 to 14km2. These results indicate that our approach holds great potential to simulate major changes in wetland water surface for ecosystem service; however, our products could capture neither the short-term water change caused by intensive rainstorm events nor the wetland change caused by human activities. [Copyright &y& Elsevier]

Abstract: The arid and semi-arid sagebrush–grass ecosystem occupies a substantial portion of rangelands in the western United States. Using remote sensing techniques to map the percent of sagebrush, grass/forb, and bare ground components is necessary for forage production estimation and natural resource management over large areas. However optical data have significant deficiencies in these ecosystems because of exposed bright soil, spectrally-indeterminate vegetation, and a large dead vegetation component. Radar data also have deficiencies caused by factors such as antenna pattern calibration, local incidence angle (LIA), soil moisture, and surface roughness. With the complementary vegetation information gained from optical data and radar data, these two datasets were fused to estimate 10-m sagebrush, grass, and bare ground percent cover in the non-forested areas of Yellowstone National Park, which is a representative native western rangeland ecosystem of the US. The datasets were processed to resolve the issues of antenna pattern calibration and LIA effect. Peak green Landsat, late fall Airborne Visible and Infrared Imaging Spectrometer (AVIRIS), and Airborne Synthetic Aperture Radar (AirSAR) data were fused in this analysis. AVIRIS, Landsat, AirSAR and elevation data were used to segment the study area into two main subcategories of “pure grass” and “mixed sagebrush and grass”. Landsat Tasseled Cap Greenness (LTCG) was used to retrieve bare land and grass percentages in pure grass areas. In the areas with mixed grass and sagebrush, standardized LTCG and radar C vv were used to derive the vegetation cover percentage, and the ratio of standardized LTCG and radar L hv was further used to calculate the relative abundance of sagebrush and grass. Comparison between the field and remote sensing estimations shows the correlation coefficients were 0.838, 0.746, and 0.830 for bare land, grass, and sagebrush, respectively. When grouped into three discrete categories of “low”, “medium”, and “high”, the overall accuracies were 79.4%, 75.9%, and 77.6%, respectively. Our study shows the potential for application of global spaceborne C- and L-band radar and optical data fusion for large-area rangeland monitoring. [Copyright &y& Elsevier]

Abstract: The Coarse Woody Debris (CWD) quantity, defined as biomass per unit area (t/ha), and the quality, defined as the proportion of standing dead logs to the total CWD quantity, greatly contribute to many ecological processes such as forest nutrient cycling, tree regeneration, wildlife habitat, fire dynamics, and carbon dynamics. However, a cost-effective and time-saving method to determine CWD is not available. Very limited literature could be found that applies remote sensing technique to CWD inventory. In this paper, we fused the wall-to-wall multi-frequency and multi-polarization Airborne Synthetic Aperture Radar (AirSAR) and optical Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) to estimate the quantity and quality of CWD in Yellowstone post-fire forest ecosystem, where the severe 1988 fire event resulted in high spatial heterogeneity of dead logs. To relate backscatter values to CWD metrics, we first reduced the terrain effect to remove the interference of topography on AirSAR backscatter. Secondly, we removed the influence of regenerating sapling by quadratic polynomial fitting between AVIRIS Enhanced Vegetation Index (EVI) and different channels backscatters. The quantity of CWD was derived from P hh and P hv, and the quality of CWD was derived from P hh aided by the ratio of L hv and P hh. Two maps of Yellowstone post-fire CWD quantity and quality were produced. The calculated CWD quantity and quality were validated by extensive field surveys. Regarding CWD quantity, the correlation coefficient between measured and predicted CWD is only 0.54 with mean absolute error up to 29.1 t/ha. However, if the CWD quantity was discretely classified into three categories of “≤60”, “60–120”, and “≥120”, the overall accuracy is 65.6%; if classified into two categories of “≤90” and “≥90”, the overall accuracy is 73.1%; if classified into two categories of “≤60” and “≥60”, the overall accuracy is 84.9%. This indicates our attempt to map CWD quantity spatially and continuously achieved partial success; however, the general and discrete categories are reasonable. Regarding CWD quality, the overall accuracy of 5 types (Type 1—standing CWD ratio ≥40%; Type 2—15%≤standing CWD ratio <40%; Type 3—7%≤standing CWD ratio<15%; Type 4—3%≤standing CWD ratio <7%; Type 5—standing CWD ratio <3%) is only 40.32%. However, when type 1, 2, 3 are combined into one category and type 4 and 5 are combined into one category, the overall accuracy is 67.74%. This indicates the partial success of our initial results to map CWD quality into detailed categories, but the result is acceptable if solely very coarse CWD quality is considered. Bias can be attributed to the complex influence of many factors, such as field survey error, sapling compensation, terrain effect reduction, surface properties, and backscatter mechanism understanding. [Copyright &y& Elsevier]

Abstract: The density of lodgepole pine (Pinus contorta) sapling regeneration was mapped in areas burned during the 1988 wildfires across Yellowstone National Park (YNP), Wyoming, USA. Hyperspectral image analysis and field measurements were combined across the entire YNP extent. Airborne Visible Infrared Imaging Spectrometer (AVIRIS) image data from 2006 were used to compute ten different vegetation indices (VI). The ten VIs were combined to build multiple regression models for predicting and mapping post-fire sapling density. Four different forms of regression modeling were applied to derive the highest possible prediction accuracy (correlation coefficient of R 2 =0.83). Pine sapling regeneration 19years after large-scale wildfires showed a high level of variability in patch density (ranging from 14/100ha to 57/100ha), whereas sapling density measured previously from the first decade following wildfire was more uniform (10/100ha to 21/100ha). The ecosystem-level clumpiness index showed major shifts in aggregation of different sapling density classes, and was consistent with an overall decrease in estimated sapling density of nearly 50% between 1998 and 2007. This analysis revealed important succession patterns and processes in post-fire forest regeneration for the Greater Yellowstone Area (GYA). [Copyright &y& Elsevier]