Estimation of forest structural and compositional variables using ALS data and multi-seasonal satellite imagery

Shang, Chaoqun; Treitz, Paul; Caspersen, John P.; Jones, Terry

doi:10.1016/j.jag.2018.10.002

Cited by 13 publications

(9 citation statements)

References 96 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We were able to significantly improve estimates of the amount, location and condition of longleaf pine and other forest ecosystems across the Fort Stewart SGA using Landsat 8 imagery and FIA field data. Similar to other studies that relate field data with remotely sensed information to estimate aspects of the forested condition, BAH tended to be more strongly correlated with spectral metrics than TPH [12,[41][42][43]. Additionally, our study also supports the idea that multi-temporal imagery and finer resolution imagery such as NAIP provides additional information over using only single season imagery when predicting forest characteristics [41,44].…”

Section: Discussionsupporting

confidence: 87%

“…Similar to other studies that relate field data with remotely sensed information to estimate aspects of the forested condition, BAH tended to be more strongly correlated with spectral metrics than TPH [12,[41][42][43]. Additionally, our study also supports the idea that multi-temporal imagery and finer resolution imagery such as NAIP provides additional information over using only single season imagery when predicting forest characteristics [41,44]. However, due to the NAIP image mosaicking process, NAIP based predictor variables provided less information than we anticipated in many of our models and also contained added spatial variability, making raster surfaces derived from those predictor variables less desirable from an applied perspective.…”

Section: Discussionsupporting

confidence: 80%

See 1 more Smart Citation

Using Forest Inventory Data with Landsat 8 Imagery to Map Longleaf Pine Forest Characteristics in Georgia, USA

et al. 2019

View full text Add to dashboard Cite

This study improved on previous efforts to map longleaf pine (Pinus palustris) over large areas in the southeastern United States of America by developing new methods that integrate forest inventory data, aerial photography and Landsat 8 imagery to model forest characteristics. Spatial, statistical and machine learning algorithms were used to relate United States Forest Service Forest Inventory and Analysis (FIA) field plot data to relatively normalized Landsat 8 imagery based texture. Modeling algorithms employed include softmax neural networks and multiple hurdle models that combine softmax neural network predictions with linear regression models to estimate key forest characteristics across 2.3 million ha in Georgia, USA. Forest metrics include forest type, basal area and stand density. Results show strong relationships between Landsat 8 imagery based texture and field data (map accuracy > 0.80; square root basal area per ha residual standard errors < 1; natural log transformed trees per ha < 1.081). Model estimates depicting spatially explicit, fine resolution raster surfaces of forest characteristics for multiple coniferous and deciduous species across the study area were created and made available to the public in an online raster database. These products can be integrated with existing tabular, vector and raster databases already being used to guide longleaf pine conservation and restoration in the region.

show abstract

Section: Discussionsupporting

confidence: 87%

Section: Discussionsupporting

confidence: 80%

Using Forest Inventory Data with Landsat 8 Imagery to Map Longleaf Pine Forest Characteristics in Georgia, USA

et al. 2019

View full text Add to dashboard Cite

show abstract

“…We set a height threshold of 1.37 m for this study, below which returns were assumed to be ground returns and ignored in metric calculation. However, smaller (i.e., younger) trees grow faster than large (i.e., older) trees and thus accumulate carbon at a faster rate [88]. The exclusion of these trees in our inventory and in subsequent modelling could be a missing component that contributes to predicted carbon accumulation (i.e., under-estimating carbon accumulation).…”

Section: Limitations and Future Workmentioning

confidence: 98%

“…Since results were similar regardless of species abundance data used and regardless of mortality, it implies that further refinement is needed in creating accurate DBH for G&Y model initialization and that perhaps using a distribution of DBH values may improve modeling results. Indeed, perhaps more information is needed that goes beyond the ability of LiDAR to provide for carbon stocks or accumulation, and that integration of more ancillary data is necessary, such as aerial photography, Sentinel-2 A or other remotely-sensed data to provide stronger initial estimates of BA_dist and SD/SDD prior to G&Y modeling [87,88].…”

Section: Assessment Of Gandy Modelsmentioning

confidence: 99%

Predicting Carbon Accumulation in Temperate Forests of Ontario, Canada Using a LiDAR-Initialized Growth-and-Yield Model

et al. 2020

View full text Add to dashboard Cite

Climate warming has led to an urgent need for improved estimates of carbon accumulation in uneven-aged, mixed temperate forests, where high uncertainty remains. We investigated the feasibility of using LiDAR-derived forest attributes to initialize a growth and yield (G&Y) model in complex stands at the Petawawa Research Forest (PRF) in eastern Ontario, Canada; i.e., can G&Y models based on LiDAR provide accurate predictions of aboveground carbon accumulation in complex forests compared to traditional inventory-based estimates? Applying a local G&Y model, we forecasted aboveground carbon stock (tons/ha) and accumulation (tons/ha/yr) using recurring plot measurements from 2012–2016, FVS1. We applied statistical predictors derived from LiDAR to predict stem density (SD), stem diameter distribution (SDD), and basal area distribution (BA_dist). These data, along with measured species abundance, were used to initialize a second model (FVS2). A third model was tested using LiDAR-initialized tree lists and photo-interpreted estimates of species abundance (i.e., FVS3). The carbon stock projections for 2016 from the inventory-based G&Y model) were equivalent to validation carbon stocks measured in 2016 at all size-class levels (p < 0.05), while LiDAR-based G&Y models were not. None of the models were equivalent to validation data for accumulation (p > 0.05). At the plot level, LiDAR-based predictions of carbon accumulation over a nine-year period did not differ when using either inventory or photo-interpreted species (p < 0.05). Using a constant mortality rate, we also found statistical equivalency of inventory and photo-interpreted accumulation models for all size classes ≥17 cm. These results suggest that more precise information is needed on tree characteristics than we could derive from LiDAR, but that plot-level species information is not as critical for predictions of carbon accumulation in mixed-species forests. Further work is needed on the use of LiDAR to quantify stand properties before this technique can be used to replace recurring plot measurements to quantify carbon accumulation.

show abstract

“…We used the multivariate implementation of RF provided by R function rfsrc and used default parameter settings [34,35]. Several authors have found that RF models built with subsets of predictors performed better than those built with all predictors (e.g., [17,18,36]), hence we tested a variable selection strategy in which forward selection in a redundancy analysis (RDA) was used to pick best subsets of predictors. This is similar to the "greedy-Wilks" strategy used by Lim et al [37].…”

Section: Model Creationmentioning

confidence: 99%

Use of Sentinel-2 Data to Improve Multivariate Tree Species Composition in a Forest Resource Inventory

et al. 2021

Self Cite

View full text Add to dashboard Cite

Aerial-photo interpreted inventories of forest resources, including tree species composition, are valuable in forest resource management, but are expensive to create and can be relatively inaccurate. Because of differences among tree species in their spectral properties and seasonal phenologies, it might be possible to improve such forest resource inventory information (FRI) by using it in concert with multispectral satellite information from multiple time periods. We used Sentinel-2 information from nine spectral bands and 12 dates within a two-year period to model multivariate percent tree species composition in >51,000 forest stands in the FRI of south-central Ontario, Canada. Accuracy of random forest (RF) and convolutional neural network (CNN) predictions were tested using species-specific basal area information from 155 0.25-ha field plots. Additionally, we created models using the Sentinel-2 information in concert with the field data and compared the accuracy of these models and the FRI-based models by use of basal areas from a second (13.7-ha) field data set. Based on average R2 values across species in the two field data sets, the Sentinel-FRI models outperformed the FRI, showing 1.5- and 1.7-fold improvements relative to the FRI for RF and 2.1- and 2.2-fold improvements for CNN (mean R2: 0.141–0.169 (FRI); 0.217–0.295 (RF); 0.307–0.352 (CNN)). Models created with the field data performed even better: improvements relative to the FRI were 2.1-fold for RF and 2.8-fold for CNN (mean R2: 0.169 (FRI); 0.356 (RF); 0.469 (CNN)). As predicted, R2 values between FRI- and field-trained predictions were higher than R2 values with the FRI. Of the 21 tree species evaluated, 8 relatively rare species had poor models in all cases. Our multivariate approach allowed us to use more FRI stands in model creation than if we had been restricted to stands dominated by single species and allowed us to map species abundances at higher resolution. It might be possible to improve models further by use of tree stem maps and incorporation of the effects of canopy disturbances.

show abstract

Estimation of forest structural and compositional variables using ALS data and multi-seasonal satellite imagery

Cited by 13 publications

References 96 publications

Using Forest Inventory Data with Landsat 8 Imagery to Map Longleaf Pine Forest Characteristics in Georgia, USA

Using Forest Inventory Data with Landsat 8 Imagery to Map Longleaf Pine Forest Characteristics in Georgia, USA

Predicting Carbon Accumulation in Temperate Forests of Ontario, Canada Using a LiDAR-Initialized Growth-and-Yield Model

Use of Sentinel-2 Data to Improve Multivariate Tree Species Composition in a Forest Resource Inventory

Contact Info

Product

Resources

About