Abstract
Reliable monitoring of the leaf area index (LAI) is required to understand and model the exchanges of carbon, water, and energy between forests and the atmosphere. Several remote sensing methods were developed to measure LAI globally, but most of them did not consider the specificities of larch forests despite the huge area such forests cover in East Siberia. We developed a satellite-based method to estimate the LAI of larch forests, separating overstory LAI (LAIo) from the understory LAI (LAIu). This method is based on the seasonal dynamics of a spectral index called NDWI, and it was developed and calibrated using three-dimensional radiative transfer simulations based on forest scenes representative of larch forest structure, with particular consideration of the typical clumped shoot structure of larch. Model-based sensitivity analyses indicated a maximum error of up to 26% under known noise levels. The method was applied to medium resolution SPOT-VEGETATION data, and was evaluated by comparing its results to global LAI products and to reference data (in situ LAI and LAI derived from high resolution Landsat ETM+ data). Our LAIo slightly underestimated field-obtained LAIo data but reproduced the spatial pattern of LAI better than CYCLOPES and MOD15 LAI products. The main drawbacks of MOD15 were unrealistically strong temporal variability and the fact that LAI began to increase earlier than the overstory leaf appearance date. In contrast, CYCLOPES LAI started increasing at the correct time and displayed smooth temporal evolutions; however, because these data are effective LAI, they may be difficult to use. At the continental scale, our total LAI is close to those from CYCLOPES and MODIS. Overall, our method provides a good alternative to MOD15 and CYCLOPES, especially for studies requiring separate estimates of LAIo and LAIu and true LAI instead of effective LAI.
Abstract from the paper’s OpenAlex record, shown for indexing and research. Copyright remains with the authors or the publisher. Read at the source ↗
The text the model read, with the source record’s inline markup (italics, sub- and superscripts) rendered and its character entities decoded; quote offsets refer to the raw text.
Claims 7
Claim · model-written summary
Model sensitivity analysis found a maximum error of up to 26%.
Verbatim
e0“Model-based sensitivity analyses indicated a maximum error of up to 26% under known noise levels.”
t0“larch”
Claim · model-written summary
The estimated overstory LAI slightly underestimated field observations.
Verbatim
e0“Our LAIo slightly underestimated field-obtained LAIo data”
t0“larch”
Claim · model-written summary
The method reproduced spatial LAI patterns better than CYCLOPES and MOD15.
Verbatim
e0“reproduced the spatial pattern of LAI better than CYCLOPES and MOD15 LAI products”
t0“larch”
Claim · model-written summary
MOD15 showed unrealistically strong temporal variability.
Verbatim
e0“The main drawbacks of MOD15 were unrealistically strong temporal variability”
t0“larch”
Claim · model-written summary
MOD15 LAI increased before overstory leaf appearance.
Verbatim
e0“LAI began to increase earlier than the overstory leaf appearance date”
t0“larch”
Claim · model-written summary
CYCLOPES LAI increased at the correct time and evolved smoothly.
Verbatim
e0“CYCLOPES LAI started increasing at the correct time and displayed smooth temporal evolutions”
t0“larch”
Claim · model-written summary
Continental total LAI from the method was close to CYCLOPES and MODIS estimates.
Verbatim
e0“At the continental scale, our total LAI is close to those from CYCLOPES and MODIS.”
t0“larch”