Using Very-High-Resolution Multispectral Classification to Estimate Savanna Fractional Vegetation Components

Andrea E Gaughan,Nicholas E Kolarik,Jonathan Salerno,Michael Drake,Joel Hartter,Lin Cassidy,Forrest R Stevens,Kyle Woodward,Narcisa G Pricope,Karen M Bailey

doi:10.3390/rs14030551

Abstract

Characterizing compositional and structural aspects of vegetation is critical to effectively assessing land function. When priorities are placed on ecological integrity, remotely sensed estimates of fractional vegetation components (FVCs) are useful for measuring landscape-level habitat structure and function. In this study, we address whether FVC estimates, stratified by dominant vegetation type, vary with different classification approaches applied to very-high-resolution small unoccupied aerial system (UAS)-derived imagery. Using Parrot Sequoia imagery, flown on a DJI Mavic Pro micro-quadcopter, we compare pixel- and segment-based random forest classifiers alongside a vegetation height-threshold model for characterizing the FVC in a southern African dryland savanna. Results show differences in agreement between each classification method, with the most disagreement in shrub-dominated sites. When compared to vegetation classes chosen by visual identification, the pixel-based random forest classifier had the highest overall agreement and was the only classifier not to differ significantly from the hand-delineated FVC estimation. However, when separating out woody biomass components of tree and shrub, the vegetation height-threshold performed better than both random-forest approaches. These findings underscore the utility and challenges represented by very-high-resolution multispectral UAS-derived data (~10 cm ground resolution) and their uses to estimate FVC. Semi-automated approaches statistically differ from by-hand estimation in most cases; however, we present insights for approaches that are applicable across varying vegetation types and structural conditions. Importantly, characterization of savanna land function cannot rely only on a “greenness” measure but also requires a structural vegetation component. Underscoring these insights is that the spatial heterogeneity of vegetation structure on the landscape broadly informs land management, from land allocation, wildlife habitat use, natural resource collection, and as an indicator of overall ecosystem function.

Highlights

Remote sensing is valuable for mapping these heterogeneous landscapes at a variety of spatiotemporal scales [8] and numerous studies highlight the relevance of object-based image analysis (OBIA) [9], pixel-based analysis [10], and/or certain machinelearning classification techniques [11,12]
We compared three different fractional vegetation components (FVCs) techniques: a pixel-based random forest (P-RF), an object- or segmentation-based random forest (S-RF), and a thresholding approach, and we evaluated their mean relative difference from a hand classification of random points
For the P-RF, the canopy height model (CHM) covariate was followed by red-edge NDVI covariates and the NIR-based NDVI metrics

Summary

Introduction

2022, 14, 551 transitional drylands [3] that vary from open grassland, characterized by tall grasses and minimal woody composition, to predominantly woody and minimal herbaceous cover [4]. Vegetation composition and structure is important as shifts along the woody-herbaceous continuum have implications for global carbon cycling, ecosystem health, and human livelihoods [7]. Remote sensing is valuable for mapping these heterogeneous landscapes at a variety of spatiotemporal scales [8] and numerous studies highlight the relevance of object-based image analysis (OBIA) [9], pixel-based analysis [10], and/or certain machinelearning classification techniques (e.g., random forest) [11,12]. This paper addresses that gap for savannas, a notably challenging landscape, where the composition and the structure of the landscape have important implications for land function

Methods

Results

Discussion

Conclusion