Drought is one of the most important natural hazards impacting ecosystem carbon cycles. However, it is challenging to quantify the impacts of drought on ecosystem carbon balance and several factors hinder our explicit understanding of the complex drought impacts. First, drought impacts can have different time dimensions such as simultaneous, cumulative, and lagged impacts on ecosystem carbon balance. Second, drought is not only a multiscale (e.g., temporal and spatial) but also a multidimensional (e.g., intensity, time‐scale, and timing) phenomenon, and ecosystem production and respiration may respond to each drought dimension differently. In this study, we conducted a comprehensive drought impact assessment on ecosystem productivity and respiration in humid regions by including different drought dimensions using global FLUXNET observations. Short‐term drought (e.g., 1‐month drought) generally did not induce a decrease in plant productivity even under high severity drought. However, ecosystem production and respiration significantly decreased as drought intensity increased for droughts longer than 1 month in duration. Drought timing was important, and ecosystem productivity was most vulnerable when drought occurred during or shortly after the peak vegetation growth. We found that lagged drought impacts more significantly affected ecosystem carbon uptake than simultaneous drought, and that ecosystem respiration was less sensitive to drought time scale than ecosystem production. Overall, our results indicated that temporally‐standardized meteorological drought indices can be used to reflect plant productivity decline, but drought timing, antecedent, and cumulative drought conditions need to be considered together.
Fog is an important water source that alleviates vegetation water stress, especially for dryland ecosystems. Comprehensive knowledge of fog and rainfall effects can help us better understand dryland vegetation responses to current and future climates. However, the differences between fog and rainfall effects on vegetation are poorly understood. This study compared the effects of fog and rainfall on vegetation greenness changes based on the ground‐level meteorological observations in the Namib Desert and the satellite vegetation index. The vegetation index and its first derivative were utilized to indicate vegetation greenness and its change rate, respectively. Results showed that fog played a more significant role than rainfall in explaining vegetation greenness change rates, while accumulated rainfall was more important than fog in determining vegetation greenness. Soil temperature was an important factor in explaining vegetation greenness changes. These findings offer key insights into how fog and rainfall differentially contribute to vegetation greenness changes.
3 October 2023 in Communications in Statistics - Simulation and Computation
Heterogeneous treatment effect estimation is an essential element in the practice of tailoring treatment to suit the characteristics of individual patients. Most existing methods are not sufficiently robust against data irregularities. To enhance the robustness of the existing methods, we recently put forward a general estimating equation that unifies many existing learners. But the performance of model-based learners depends heavily on the correctness of the underlying treatment effect model. This paper addresses this vulnerability by converting the treatment effect estimation to a weighted supervised learning problem. We combine the general estimating equation with supervised learning algorithms, such as the gradient boosting machine, random forest, and artificial neural network, with appropriate modifications. This extension retains the estimators' robustness while enhancing their flexibility and scalability. Simulation shows that the algorithm-based estimation methods outperform their model-based counterparts in the presence of nonlinearity and non-additivity. We developed an R package, RCATE, for public access to the proposed methods. To illustrate the methods, we present a real data example to compare the blood pressure-lowering effects of two classes of antihypertensive agents.