Third, novel fusion algorithms can help to improve the spatial/temporal coverage and the accuracy of global reservoir products. As mentioned in sections 3.2.2 and 3.2.3, data fusion techniques have been used in a number of studies. These algorithms can be classified into two groups, the first of which focuses on fusing measurements of a single variable from different data sources. For elevation, the most common practice involves adopting data from all radar altimeters (Crétaux et al., 2015). For area, Zhang & Gao (2016) combined Advanced Microwave Scanning Radiometer and Earth Observing System (AMSR‐E) with MODIS observations to estimate reservoir area under all‐weather conditions. High‐resolution Landsat observations are also commonly used for downscaling the medium‐resolution MODIS results (Che et al., 2015). The second group is focused on taking advantage of elevation‐area‐storage relationships, such that a single variable with good accuracy (and/or good spatial/temporal coverage) can be used for inferring the values of the others (Gao et al., 2012). Powered by the ever increasing cloud‐computing platforms (such as the Google Earth Engine), the new generation of fusion algorithms will be able to incorporate elements from both groups.
Last, recent developments of the water management component used in earth system modeling tools has brought forth great promise for the investigation of human intervention in a holistic manner (Li et al., 2015; Voisin et al., 2017; Yigzaw et al., 2018). As such, the assimilation of remote sensing reservoir data is expected to be feasible, similar to the improved drought monitoring which has occurred using GRACE data incorporated into NLDAS (Kumar et al., 2016).
The pressing research questions related to reservoirs in the coming decades are:
1 What are the impacts of reservoir impoundments on the spatial and temporal distributions of the hydrological cycle?
2 How does reservoir storage respond to climate variability, climate change, extreme events, and human activities across scales?
3 How do we improve reservoir water management under the stress of future climate change and population growth?
To solve these questions, the science communities from different areas and disciplines need to collaborate for convergence research. For instance, remote sensing and modeling methods should be fully integrated, while the hydrology and water resources management communities should interact with decision makers more proactively.
3.5. DISCUSSION AND CONCLUSIONS
Remotely sensed reservoir data have the potential of being used with other drought indicators jointly to promote improved drought mitigation.
First, by combining the remotely sensed reservoir storage information with GRACE terrestrial water storage anomalies, optimal use of surface water and groundwater in combination would be feasible. In light of the ever‐increasing water needs and the uncertain supplies, leveraging water resources and increasing water use efficiency, including both surface water and groundwater, will be key for sustaining population growth and agricultural production in the 21st century. However, the existing GRACE‐based drought indicators examine only the total storage anomalies (Long et al., 2013; Thomas et al., 2014). A recent study has found that the drought index based on reservoir area can help fill the information gap between streamflow/runoff‐based and groundwater‐based drought indices (Zhao & Gao, 2019b). By integrating a surface water storage component, the joint management of surface water and groundwater can be achieved. For instance, Tang et al. (2010) have shown that subtracting reservoir storage changes from GRACE water storage anomalies could provide better support for water resources management purposes.
Second, a holistic examination of meteorological, agricultural, and hydrological drought indicators is crucial for purposes of mitigating socioeconomic drought. With reservoir inflow driven by precipitation and release determined by water demand, a reservoir drought index can meaningfully link meteorological drought with socioeconomic drought. It is worth noting that a given reservoir drought index is a function of inflow, evaporation, and release. While inflow and evaporation are primarily affected by meteorological drought, release is primarily used for flood control and/or meeting downstream water demand. For reservoir operation optimization under drought conditions, the trade‐offs among agricultural, industrial, and municipal supplies are complicated and are closely tied to the socioeconomic benefits. In this sense, reservoir drought indices are hybrid in nature and uniquely different from the other types of indices. By comparing meteorological drought indices with given reservoir drought indices, the impacts on reservoir storage from natural drought and human usage can be partitioned. Under the pressing situation of nonstationarity (Milly et al., 2008), a decreasing trend of reservoir storage/area/elevation could be triggered by the combination of drought and increased water use (e.g., from the growing population).
In summary, several conclusions can be drawn from this chapter. (a) Remote sensing of reservoir data offers a unique promise for monitoring hydrological drought from space. Due to the shortage of streamflow and reservoir measurements, satellite observations of surface water are invaluable. (b) Reservoir storage is deemed to be the best indicator of drought severity, as compared to elevation and area. However, such data are available only for a small portion of reservoirs globally, limited by the availability of altimetry measurements. (c) Reservoir area data, which have better spatial and temporal coverage than elevation/storage data, can be used to derive a drought index suitable for monitoring purposes at local and regional scales. (d) By fusing observations collected by past, current, and future satellite missions, storage monitoring of most global reservoirs will be feasible. (e) The skills of hydrological‐modeling‐based drought monitors can be enhanced by incorporating remotely sensed reservoir information.
ACKNOWLEDGMENTS
This chapter was supported by NASA Grant 80NSSC18K0939 to Texas A&M University.
REFERENCES
1 AMS (2004). AMS statement on meteorological drought Bulletin of the American Meteorological Society, 85, 771–773.
2 Biancamaria, S., Lettenmaier, D.P., & Pavelsky, T.M. (2016). The SWOT mission and its capabilities for land hydrology. In Cazenave, A., Champollion, N., Benveniste, J., Chen, J. (Eds.), Remote Sensing and Water Resources (pp. 117–147). Springer.
3 Bierkens, M.F. (2015). Global hydrology 2015: State, trends, and directions. Water Resources Research, 51(7), 4923–4947.
4 Birkett, C. (1994). Radar altimetry: a new concept in monitoring lake level changes. Eos, Transactions American Geophysical Union, 75(24), 273–275.
5 Birkett, C. (2000). Synergistic remote sensing of Lake Chad: Variability of basin inundation. Remote sensing of environment, 72(2), 218–236.
6 Bondeau, A., Smith, P.C., Zaehle, S., Schaphoff, S., Lucht, W., Cramer, W., et al. (2007). Modelling the role of agriculture for the 20th century global terrestrial carbon balance. Global Change Biology, 13(3), 679–706.
7 Bryant, R.G. (1999). Application of AVHRR to monitoring a climatically sensitive playa. Case study: Chott el Djerid, southern Tunisia. Earth Surface Processes and Landforms, 24(4), 283–302.
8 Busker, T., de Roo, A., Gelati, E., Schwatke, C., Adamovic, M., Bisselink, B., et al. (2018). A global lake and reservoir volume analysis using a surface