The Earth's climate is changing rapidly and unexpectedly, causing more frequent, longer and more severe droughts, with lasting impacts on plants, ecosystems, communities and people. Consequently, this is leading to an increased importance of monitoring the climate and water storage trends in different regions. This information on a global scale is already commonly derived using satellite-based geodetic techniques such as the Global Positioning System (GPS) and the Gravity Recovery and Climate Experiment (GRACE). The use of both techniques has significant advantages, especially in regions where changes in the hydrosphere are notable, such as the Amazon basin, where 25 GPS stations were lately classified as benchmarks for hydrogeodesy. We show that the vertical displacements obtained from GPS and GRACE have good spatio-temporal agreement with the Standardized Precipitation and Standardized Precipitation Evapotranspiration indices, abbreviated respectively as SPI and SPEI, for all these stations. Drought severity index (DSI) estimated separately from GPS-observed and GRACE-derived vertical displacements on a station-by-station basis is capable to identify dry and wet events previously reported for the Amazon basin. However, due to the weaknesses of both techniques, such as technique-related systematic errors or coarse spatial resolution, a few extreme hydrological events may not be properly captured by GPS-DSI and/or GRACE-DSI. To take full advantage of both techniques and overcome their weaknesses, we introduce a completely new methodology to combine individual GPS-DSI and GRACE-DSI indices. As a novelty, both indices are estimated using short-term changes (<9 months) of monthly vertical displacements observed by GPS permanent stations and those derived by GRACE for GPS locations. Then, to capture and detect drought events that either both geodetic techniques metrics missed or incorrectly depicted, the Multivariate Drought Severity Index (MDSI) is estimated through the concept of Frank copulas. We demonstrate that the MDSI captures more hydroclimatic events reported in previous studies, which are not identified by individual series of GPS-DSI or GRACE-DSI indices, and is temporally consistent with Standardized Streamflow Index (SSI) based on the in-situ river discharge changes.

中文翻译：

---

用于识别短期水文信号的新多变量干旱严重程度指数：亚马逊河流域案例研究


地球的气候正在迅速而出乎意料地变化，导致更频繁、更持久和更严重的干旱，对植物、生态系统、社区和人类产生持久影响。因此，这导致监测不同地区的气候和水储存趋势的重要性增加。全球范围内的这些信息通常使用基于卫星的大地测量技术（例如全球定位系统 （GPS） 和重力恢复和气候实验 （GRACE））来获取。使用这两种技术具有显著的优势，尤其是在水圈变化明显的地区，例如亚马逊盆地，那里的 25 个 GPS 站最近被归类为水文测地测量的基准。结果表明，GPS 和 GRACE 获得的垂直位移与所有这些站点的标准化降水和标准化降水蒸散指数（分别缩写为 SPI 和 SPEI）具有良好的时空一致性。根据 GPS 观测和 GRACE 衍生的垂直位移逐站分别估计的干旱严重程度指数 （DSI） 能够识别以前为亚马逊盆地报告的干湿事件。然而，由于这两种技术的弱点，例如与技术相关的系统误差或粗略的空间分辨率，GPS-DSI 和/或 GRACE-DSI 可能无法正确捕获一些极端水文事件。为了充分利用这两种技术并克服它们的弱点，我们引入了一种全新的方法来组合单独的 GPS-DSI 和 GRACE-DSI 指数。 作为一个新颖性，这两个指数都是使用 GPS 永久站观测到的每月垂直位移的短期变化（<9 个月）和 GRACE 得出的 GPS 位置的变化来估计的。然后，为了捕获和检测两种大地测量技术指标都遗漏或错误描述的干旱事件，通过 Frank copulas 的概念估计多变量干旱严重程度指数 （MDSI）。我们证明 MDSI 捕获了以前研究中报告的更多水文气候事件，这些事件不是通过单个系列的 GPS-DSI 或 GRACE-DSI 指数来识别的，并且在时间上与基于原位河流流量变化的标准化流速指数 （SSI） 一致。