Vegetation–soil moisture coupling metrics from dual-polarization microwave radiometry using regularization

Abstract

Soil and vegetation water content are closely coupled via complex physiological and ecohydrological processes. Joint passive microwave retrievals of soil moisture θ and vegetation optical depth τ potentially provide unparalleled insight into these couplings on a global scale. However, this requires careful data analyses. Using a novel coupling distortion metric Rs2, we show that snapshot dual-polarization retrievals of τ and θ– widely used in vegetation studies – are spuriously correlated for SMAP L-band observations. Naive estimates of τ–θ coupling metrics are thus grossly distorted across a range of time scales. To mitigate the spurious correlations, we introduce a regularized retrieval algorithm. Our regularization algorithm exploits the assumed slowly changing nature of τ by penalizing rapid variations in the τ estimates. The degree of regularization r must balance a trade-off, as we find overregularization due to the oversmoothing of τ also distorts coupling estimates. When r is chosen to balance the trade-off according to Rs2, the spurious correlations are found to essentially vanish. The estimates of τ–θ correlation change substantially compared to non-regularized retrievals. The changes are largest (∼0.5) over high-biomass at time scales of up to two weeks, but sizeable differences are also found on longer time scales. Our analyses show that estimating vegetation–soil moisture coupling metrics benefits from dedicated retrievals and data analysis approaches. Provided the uncertainties are carefully accounted for, satellite radiometry offers exciting opportunities to study ecohydrological interactions such as plant water uptake and hydraulics.

Plant phenotyping relevance

植物光学的厚さτ（植生水分状態の指標）を衛星マイクロ波から推定する正則化検索アルゴリズムと結合指標を開発・評価しており、植生状態の取得手法が研究の中心です。

Code and data availability

公開本文の所在を確認できませんでした。非公開または購読が必要な可能性があります。

No evidence-backed public reproduction asset is currently recorded.