In its three-dimensional (3-D) characterization, drought is approached as an event whose spatial extent changes over time. Each drought event has an onset and end time, a location, a magnitude, and a spatial trajectory. These characteristics help to analyze and describe how drought develops in space and time, i.e., drought dynamics. Methodologies for 3-D characterization of drought include a 3-D clustering technique to extract the drought events from the hydrometeorological data. The application of the clustering method yields small ‘artifact’ droughts. These small clusters are removed from the analysis with the use of a cluster size filter. However, according to the literature, the filter parameters are usually set arbitrarily, so this study concentrated on a method to calculate the optimal cluster size filter for the 3-D characterization of drought. The effect of different drought indicator thresholds to calculate drought is also analyzed. The approach was tested in South America with data from the Latin American Flood and Drought Monitor (LAFDM) for 1950–2017. Analysis of the spatial trajectories and characteristics of the most extreme droughts is also included. Calculated droughts are compared with information reported at a country scale and a reasonably good match is found.

Conceptual hydrological models imply a simplification of the complexity of the hydrological system; however, it lacks the flexibility in reproducing a wide range of the catchment responses. Usually, a trade-off is done to sacrifice the accuracy of a specific aspect of the system behavior in favor of the accuracy of other aspects. This study evaluates the benefit of using a modular approach, “The fuzzy committee model” of building specialized models (same structure associated with different parameter realization) to reproduce specific responses (high and low flow response) of the catchment. The study also assesses the applicability of using predicted runoff from both specialized models with certain weights based on a fuzzy membership function to form a fuzzy committee model. This research continues to explore the fuzzy committee models first presented by [Solomatine, 2006] and further developed by [Fenicia et al., 2007; Kayastha et al., 2013]. In this paper, weighting schemes with power parameter values are investigated. A thorough study is conducted on the relation between the fuzzy committee variables (the membership functions and the weighting schemes), and their effect on the model performance. Furthermore, the Fuzzy committee concept is applied on a conceptual distributed model with two cases, the first with lumped catchment parameters and the latter with distributed parameters. A comparison between different combinations of the fuzzy committee variables showed the superiority of all Fuzzy Committee models over single models. Fuzzy committee of distributed models performed well, especially in capturing the highest peak in the calibration data set; however, it needs further study of the effect of model parameterization on the model performance and uncertainty.

Drought is a major threat to global agriculture and can trigger or intensify food price increase and migration. Assessment and monitoring are essential for proper drought management. Drought indices play a fundamental task in this respect. This research introduces the Wet-environment Evapotranspiration and Precipitation Standardized Index (WEPSI) for drought assessment and monitoring. WEPSI is inspired by the Standardized Precipitation Evapotranspiration Index (SPEI), in which water supply and demand are incorporated into the drought index calculation. WEPSI considers precipitation (P) for water supply and wet-environment evapotranspiration (ETw) for water demand. We use an asymmetric complementary relationship to calculate ETw using actual (ETa) and potential evapotranspiration (ETp). WEPSI is tested in the transboundary Lempa River basin located in the Central American dry corridor. ETw is estimated based on evapotranspiration data calculated using the Water Evaluation And Planning (WEAP) system hydrological model. To investigate the performance of our introduced drought index, we compare it with two well-known meteorological indices (Standardized Precipitation Index and SPEI), together with a hydrological index (Standardized Runoff Index), in terms of correlation and mutual information (MI). We also compare drought calculated with WEPSI and historical information, including crop cereal production and Oceanic Niño Index (ONI) data. The results show that WEPSI has the highest correlation and MI compared with the three other indices used. It is also consistent with the records of crop cereal production and ONI. These findings show that WEPSI can be applied for agricultural drought assessments.

Finding optimum balances between conflicting interests in multipurpose reservoirs often represents an important challenge for decision makers. This study assesses the use of different computational tools to obtain optimal reservoir operations applied to the Hatillo dam in the Dominican Republic. A multiobjective optimization approach is used, in which non-dominated sorting genetic algorithm II (NSGAII) and multi-objective evolutionary algorithm based on decomposition (MOEA/D) optimizers are applied to models that simulate reservoir operations. Three different Machine Learning (ML) models, namely, the multilayer perceptron (MLP), the radial basis network (RBN) and the linear function (LF), are employed to learn the current operation of the system. Subsequently, a general model is proposed to simulate daily reservoir operations (2009-2019), integrating water balances, physical constraints of the dam components and the ML models, the latter defining daily controlled discharges. In the optimization process, the ML parameters are the decision variables, while the objectives evaluated are irrigation, hydropower generation and flood control. The results are compared with the actual operation of the reservoir. Three dimensional Pareto fronts are obtained, from which, the wide variety of operations can be evidenced. The flood control objective was found to have a wide room for improvement over the current operation of the reservoir, and several of the solutions found improve the current operation for the three proposed objectives. The MLP models tend to generate the best results for this case study and the NSGAII optimizer generates the best optimization results.

Extreme runoff modeling is hindered by the lack of sufficient and relevant ground information and the low reliability of physically-based models. The authors propose to combine precipitation Remote Sensing (RS) products, Machine Learning (ML) modeling, and hydrometeorological knowledge to improve extreme runoff modeling. The approach applied to improve the representation of precipitation is the object-based Connected Component Analysis (CCA), a method that enables classifying and associating precipitation with extreme runoff events. Random Forest (RF) is employed as a ML model. We used 2.5 years of nearly-real-time hourly RS precipitation from the PERSIANN-CCS and IMERG-early run databases (spatial resolutions of 0.04 o and 0.1 o , respectively), and runoff at the outlet of a 3391 km 2-basin located in the tropical Andes of Ecuador. The developed models show the ability to simulate extreme runoff for the cases of long-duration precipitation events regardless of the spatial extent, obtaining Nash-Sutcliffe efficiencies (NSE) above 0.72. On the contrary, we found an unacceptable model performance for a combination of short-duration and spatially-extensive precipitation events. The strengths/weaknesses of the developed ML models are attributed to the ability/difficulty to represents complex precipitation-runoff responses.