Separating of Overstory and Understory Leaf Area Indices for Global Needleleaf and Deciduous Broadleaf Forests by Fusion of MODIS and MISR Data

Abstract

<p><strong>Abstract.</strong> Forest overstory and understory layers differ in carbon and water cycle regimes, phenology, as well as ecosystem functions. Separate retrievals of Leaf Area Index (LAI) for these two layers would help to improve modeling forest biogeochemical cycles and evaluating forest ecosystem functions. In this paper, overstory and understory LAI values were estimated separately for global needleleaf and deciduous broadleaf forests by fusing MISR and MODIS observations. Monthly forest understory LAI was retrieved from the forest understory reflectivity estimated using MISR data. After correcting for the background contribution using monthly mean forest understory reflectivities, the forest overstory LAI was estimated from MODIS observations. The results demonstrate that forest understory vegetation is mainly distributed in the boreal forest zones at northern latitudes. Significant seasonal variations are presented for understory vegetation in these zones with LAI values up to 2–3 from June to August. Higher understory LAI values are found in needleleaf forests (with mean value of 1.06 for evergreen needleleaf forests and 1.04 for deciduous needleleaf forests) than in deciduous broadleaf forests (0.96) due to the more clumped foliage and easier penetration of light to the forest floor in needleleaf forests. The magnitude of seasonal variations of overstory LAI is larger for deciduous needleleaf forests than those of evergreen and deciduous needleleaf forests. In contrast, for forest understory, needleleaf forests show larger seasonal variations than broadleaf forests. Spatially and seasonally variable forest understory reflectivity helps to account for the effects of the forest background on LAI retrieval while compared with constant forest background. The retrieved forest overstory and understory LAI values were compared with an existing dataset for larch forests in Northern Eurasia. The retrieved overstory and understory LAI is close to that of the existing dataset, with an absolute error of 0.34 (0.06), relative error of 21.1 % (14.3 %) and RMSE of 0.93 (0.29) for overstory (understory). The comparisons between our results and field measurements in eight forest sites show that the R<sup>2</sup> are 0.62 and 0.52, and the RMSE are 0.62 and 1.36 for understory and overstory LAI, respectively.</p>

Estimation of Woody and Herbaceous Leaf Area Index in Sub‐Saharan Africa Using MODIS Data

“…In this paper, we have presented an approach for separating MODIS LAI into woody and herbaceous constituents across Sub‐Saharan Africa (SSA) where ~65% of terrestrial biomes are savanna ecosystems with woody‐herbaceous mixtures (Archibold, ). While several authors have developed methods for LAI partitioning in temperate and boreal regions (Huang et al, ; Kobayashi et al, ; Liu et al, ), and woody cover separation from herbaceous cover at local to regional scales in Africa (Brandt et al, ; Gessner et al, ; Wagenseil & Samimi, ), our analysis is unique in providing long‐term woody and herbaceous LAI phenologies for tropical Africa. Separated woody and herbaceous LAI allow users to understand the separate and distinct phenology and function of woody and herbaceous vegetation in ecosystem processes across SSA.…”

“…Attempts have been made to partition remote sensing products into separate woody and herbaceous components. Liu et al () partitioned understory and overstory LAI in temperate and boreal forests, using Moderate Resolution Imaging Spectroradiometer (MODIS) and Multi‐angle Imaging Spectroradiometer (MISR) data sets. However, their method is dependent on the ability to capture realistic forest background reflectivities, and neither tropical savannas nor moist tropical forests were included in their analysis.…”

An Overview of Global Leaf Area Index (LAI): Methods, Products, Validation, and Applications

“…A clear distinction between the various LAI definitions (Table ) is desirable in validation studies since most remote sensing products represent the true LAI (Table ). Separate consideration of the overstory and understory LAI would enable more efficient validation of the storied LAI products (e.g., Y. Liu, Liu, et al, ) and the LiDAR‐derived LAI vertical profiles (e.g., H. Tang et al, ). Over these sites, high‐resolution reference data should be generated, using a standardized approach with traceable results and well‐calibrated quality information.…”

Stratifying Forest Overstory and Understory for 3-D Segmentation Using Terrestrial Laser Scanning Data

“…Cabo et al [39] approximated the individual tree crown boundaries with the Voronoi diagram [40], but they failed to define the real boundaries of tree crowns accurately. However, it remains challenging to segment tree crowns using TLS data due to the complexity of vertical forest structure (overstory-understory) [41]- [43], species differences [44]- [46], and understory abundance [42]. Few studies have considered the existence of understory when extracting trees based on TLS data.…”