Research papersA satellite-based approach to estimating spatially distributed groundwater recharge rates in a tropical wet sedimentary region despite cloudy conditions
Graphical abstract
Introduction
Understanding the factors constraining groundwater recharge (GWR) is important for management and planning purposes of this water resource that is only slowly renewed (Cuthbert et al., 2019). In some regions, for instance, the abstracted groundwater over the past decades are taken from non-renewable groundwater (Döll et al., 2014), which increases, even more, the need for a better understanding of such factors. These abstractions need to be regionally regulated (Aeschbach-Hertig and Gleeson, 2012), since groundwater serves as the key strategic reserve for supplying water to societies during long-lasting droughts (Famiglietti, 2014). Such regulation, in turn, requires accurate information about the spatiotemporal distribution of natural GWR rates (Jasechko et al., 2014), including their variability and uncertainty in estimations, which are strongly sensitive to climate forcing factors, land uses and covers, watershed geomorphology and local hydrogeology (Moeck et al., 2020).
Since GWR is a key component used in many hydrological models to assess groundwater resource worldwide (Graaf et al., 2017, Wada et al., 2010), its accurate estimation constitutes a priority for stakeholders and a research challenge for the scientific community (Jasechko et al., 2014, Mohan et al., 2018). Many methods have been developed to estimate natural GWR at various spatiotemporal scales, with a wide range of complexity (Walker et al., 2019), given that GWR cannot be directly measured (Melo et al., 2015). Making use of these methods often depends on data availability, desired spatiotemporal resolution, and result representations (Walker et al., 2019).
The following five methods are commonly used to estimate GWR: 1) tracer techniques, which estimate aquifer renewal via substances in the water or specific concentrations of chemical elements, such as the chloride mass-balance method (e.g., Brunner et al., 2004, Hornero et al., 2016); 2) groundwater level monitoring in unconfined aquifers, which include examples such as water table fluctuation method (e.g., Cai and Ofterdinger, 2016, Wendland et al., 2007) and cumulative rainfall departure methods (e.g., Ahmadi et al., 2015, Weber and Stewart, 2004); 3) Darcy’s law application, which allows calculating the velocity of soil water percolation and requires knowledge of hydraulic gradient and vertical hydraulic conductivity (e.g., Callahan et al., 2012, Yin et al., 2011); 4) numerical modelling, which consists of a mathematical representation of the GWR process (e.g., Melo et al., 2015, Melo and Wendland, 2017); and 5) the water balance method, which considers the main variables of the hydrological cycle as inputs and outputs of the system (e.g., Hornero et al., 2016, Wendland et al., 2007).
Most of the aforementioned methods are based on point-scale observations (e.g., meteorological stations or boreholes), which may cause serious issues when spatial variability in the regions of concern is great (e.g., Melo and Wendland, 2017). Although such a problem can be simply ignored for regions with extremely dense observation networks, it remains persistent in most regions worldwide, especially in developing countries. For instance, in Brazil, the national ground-based monitoring network consists of about 400 wells distributed over the country, complemented by a small number of observation wells monitored in only 21 active experimental basins (Melo et al., 2020). Therefore, the chief challenge for many hydrologists is to find and utilise alternative sources of data to estimate the spatial information of GWR (Brunner et al., 2007).
The use of cutting-edge satellite-derived remote sensing technology has played a crucial role in assimilating valuable distributed observation and in modelling water resources, which would otherwise be impossible with relatively sparse ground-based measurements alone (Famiglietti et al., 2015). However, the remote sensing contributions are rather inconsistent at quantifying and estimating GWR because all current data from satellite data can only detect patterns and processes related to water resources on and above the surface (Brunner et al., 2007, Coelho et al., 2017, Lucas et al., 2015). Satellite-based observations of time-variable gravity, such as the joint mission of the Gravity Recovery and Climate Experiment (GRACE), are sensitive to variations of terrestrial water storage, including the groundwater storage changes (Tapley et al., 2004, Vasco et al., 2019, Wahr et al., 2004). Unfortunately, the low spatial resolution of GRACE-derived data limits its ability to provide localised groundwater information at an appropriate scale (Alley and Konikow, 2015, Lakshmi et al., 2018). Thus, an innovative use of satellite data to estimate GWR at local and regional scales has been recently proposed, where most of data are applied to a simplified water budget approach that uses precipitation and evapotranspiration products (e.g., Crosbie et al., 2015, Gokmen et al., 2013, Lucas et al., 2015, Műnch et al., 2013, Szilágyi et al., 2012, Szilagyi et al., 2011). This approach disregards other water balance components, such as surface runoff and soil water storage changes, which could considerably alter the estimation accuracy of GWR in some regions for short time scales (e.g., monthly).
In this context, some studies also have considered uniform surface runoff (Khalaf and Donoghue, 2012), as well as spatially distributed information about surface runoff (Coelho et al., 2017) and irrigation (Usman et al., 2015). The aforementioned studies used different remote sensing products and algorithms, but all of them were developed in regions with arid, semiarid, continental or Mediterranean climate conditions where the cloud cover is limited (Coelho et al., 2017). For some tropical regions such as Brazil, the estimation of GWR using this approach remains challenging, mainly because of the difficulties in obtaining continuous information of actual evapotranspiration data by remote sensing without substantial cloud cover. In parallel, soil moisture information from satellite observations is currently available at the global scale and can provide valuable data to update the water budget approach with information regarding water storage changes in unsaturated soil layers (Reichle et al., 2018). Accounting for this component is particularly important for understanding GWR in sedimentary aquifers, where the unsaturated vadose zone width may vary from thin to thick soil layers (Rossetti et al., 2012). Unfortunately, some satellite-based datasets are only recently available, but some applications require earlier data.
Based on this information, this study develops an innovative water budget method using satellite-based data for estimating natural spatially distributed GWR rates at annual and monthly scales in tropical wet sedimentary regions, taking into account cloudy conditions. Accordingly, this study hypothesizes that such an approach enables local and regional scale perspectives in ungauged tropical wet regions. The general and transferable strategy would be relevant to account for 1) the substantial cloud cover and 2) the water storage changes in sedimentary regions prone to monthly variations. The method also utilises spatially distributed information on precipitation and surface runoff estimated from satellite products. The major limitation of this residual approach is that the accuracy of the GWR depends on the accuracy of the other components considered in the water balance (Scanlon et al., 2002), i.e., its application is appropriated when the errors of these components are small relative to the water flux. This limitation, when a satellite-based approach is considered, is mainly identified in regions that present ground-truth measurements discrepant with the estimated products used in the water balance, especially the main input (precipitation) and output (evapotranspiration) of the system. On the other hand, ground-based evaluations are punctual and representative of small areas, hardly integrating the spatial heterogeneity of meteorological processes, especially in urban areas (Maier et al., 2020). This study used ground-truth measurements to assess the two main estimated components of the water balance (i.e., precipitation and evapotranspiration) and the GWR rates.
Section snippets
Study area
The study was carried out over an area of 1032 km2 in João Pessoa (JPA) (Paraíba, NE Brazil), which includes the metropolitan region and surrounding rural areas (Fig. 1). It consists in 1) the Gramame river basin (589.1 km2; 57.1% of the area), and 2) the right bank of the Baixo Paraíba river basin (442.9 km2; 42.9% of the area). The main source of water of the JPA metropolitan region (∼1 million inhabitants) is the Gramame-Mamuaba reservoir, with maximum volume capacity of 56.9 hm3. The water
Precipitation
The first set of analyses assessed the spatiotemporal distributions of the main input of the water balance (i.e., P) obtained from IMERG-C data and compared these with the ground-based interpolated data (henceforth Gauge) (Fig. 3). Annual P based on IMERG-C data gradually decreased from east to west, varying from 1120 to 1600 mm in 2016 and 1050 to 2300 mm in 2017. The maximum P based on Gauge observations was 1630 mm in 2016 and 2070 mm in 2017. Similarly, (Lu et al., 2019) showed consistent
Summary and conclusions
This study developed and evaluated an innovative satellite-based approach based on the water budget equation to estimate the natural GWR over by only using freely available satellite-based data. The proposed distinctive features include the capacity to address 1) ET estimations (MOD16 algorithm) in tropical wet regions frequently overlaid by substantial cloud cover and 2) water storage change estimation in the root zone (SPL4SMAU product) in sedimentary regions seasonably prone to monthly
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
The authors would like to acknowledge the financial support granted by 1) the Research Support Foundation of Paraíba State (FAPESQ-PB) (Grant REF: 88887.142311/2017-00), which also funded the contribution from Yunqing Xuan, supported in partnership with the Newton Fund, via CONFAP – The UK Academies Research Mobility 2017/2018 (Grant REF: 039/2018); 2) the Brazilian Coordination for Improvement of Higher Education Personnel (CAPES) – Finance Code 001 (Grant REF: 88887.161412/2017-00); 3) the
References (101)
- et al.
A remote sensing surface energy balance algorithm for land (SEBAL), Part 1: formulation
J. Hydrol.
(1998) - et al.
Using remote sensing to regionalize local precipitation recharge rates obtained from the Chloride Method
J. Hydrol.
(2004) - et al.
Analysis of groundwater-level response to rainfall and estimation of annual recharge in fractured hard rock aquifers, NW Ireland
J. Hydrol.
(2016) - et al.
Alluvial groundwater recharge estimation in semi-arid environment using remotely sensed data
J. Hydrol.
(2017) - et al.
Estimation and comparision of curve numbers based on dynamic land use land cover change, observed rainfall-runoff data and land slope
J. Hydrol.
(2013) - et al.
Grid box-level evaluation of IMERG over Brazil at various space and time scales
Atmos. Res.
(2019) - et al.
Integrating soil water and tracer balances, numerical modelling and GIS tools to estimate regional groundwater recharge: application to the Alcadozo Aquifer System (SE Spain)
Sci. Total Environ.
(2016) A soil-adjusted vegetation index (SAVI)
Remote Sens. Environ.
(1988)- et al.
Development of a two-band enhanced vegetation index without a blue band
Remote Sens. Environ.
(2008) - et al.
Estimating recharge distribution using remote sensing: A case study from the West Bank
J. Hydrol.
(2012)
A comparative study of available water in the major river basins of the world
J. Hydrol.
Reverification of antecedent moisture condition dependent runoff curve number formulae using experimental data of Indian watersheds
Catena
Correcting GPM IMERG precipitation data over the Tianshan Mountains in China
J. Hydrol.
Soil moisture data as a constraint for groundwater recharge estimation
J. Hydrol.
Impact of urbanization on groundwater recharge rates in Dübendorf, Switzerland
J. Hydrol.
Development of a global evapotranspiration algorithm based on MODIS and global meteorology data
Remote Sens. Environ.
Improvements to a MODIS global terrestrial evapotranspiration algorithm
Remote Sens. Environ.
Late Quaternary sedimentation in the Paraíba Basin, Northeastern Brazil: landform, sea level and tectonics in Eastern South America passive margin
Palaeogeogr. Palaeoclimatol. Palaeoecol.
Ground validation of space-borne satellite rainfall products in Malaysia
Adv. Space Res.
Research on the SCS-CN initial abstraction ratio using rainfall-runoff event analysis in the Three Gorges Area, China
CATENA
Scale dynamics of the HIDROPIXEL high-resolution DEM-based distributed hydrologic modeling approach
Environ. Modell. Software
Water balance in the Guarani Aquifer outcrop zone based on hydrogeologic monitoring
J. Hydrol.
Regional strategies for the accelerating global problem of groundwater depletion
Nat. Geosci.
Mapping groundwater recharge areas using CRD and RIB methods in the semi-arid Neishaboor Plain, Iran
Arab. J. Geosci.
Quantifying Excess Stormwater Using SCS-CN–Based Rainfall Runoff Models and Different Curve Number Determination Methods
J. Irrig. Drain. Eng.
Crop evapotranspiration - Guidelines for computing crop water requirements
Köppen’s climate classification map for Brazil
Meteorol. Z.
Climatological behavior of precipitating clouds in the northeast region of Brazil
Adv. Meteorol.
Discussion of “Interfacing the geographic information system, remote sensing, and the soil conservation service-curve number method to estimate curve number and runoff volume in the ASIR region of Saudi Arabia” by Fawzi S. Mohammad, Jan Adamowski
Arab. J. Geosci.
How can remote sensing contribute in groundwater modeling?
Hydrogeol. J.
Estimating groundwater recharge in lowland watersheds
Hydrol. Process.
Ground truthing groundwater-recharge estimates derived from remotely sensed evapotranspiration: a case in South Australia
Hydrogeol. J.
Constraining the magnitude and uncertainty of specific yield for use in the water table fluctuation method of estimating recharge
Water Resour. Res.
Observed controls on resilience of groundwater to climate variability in sub-Saharan Africa
Nature
Global-scale assessment of groundwater depletion and related groundwater abstractions: combining hydrological modeling with information from well observations and GRACE satellites
Water Resour. Res.
Sources of uncertainty in the NRCS CN model: recognition and solutions
Hydrol. Process.
The global groundwater crisis
Nat. Clim. Change
Satellites provide the big picture
Science
Assessing groundwater storage changes using remote sensing-based evapotranspiration and precipitation at a large semiarid basin scale
J. Hydrometeorol.
Advances in Water Resources A global-scale two-layer transient groundwater model: development and application to groundwater depletion
Adv. Water Resour.
Runoff probability, storm depth, and curve numbers
J. Irrig. Drainage Eng.
Using groundwater levels to estimate recharge
Hydrogeol. J.
The pronounced seasonality of global groundwater recharge
Water Resour. Res.
Evaluation of the Soil Conservation Service curve number methodology using data from agricultural plots
Hydrogeol. J.
Comparison and bias correction of TMPA precipitation products over the Lower Part of Red-Thai Binh river basin of Vietnam
Remote Sens.
Estimates of future discharges of the river Rhine using two scenario methodologies: direct versus delta approach
Hydrol. Earth Syst. Sci.
Cited by (0)
- 1
These authors contributed equally to this work.