Elsevier

Journal of Hydrology

Volume 607, April 2022, 127503
Journal of Hydrology

Research papers
A satellite-based approach to estimating spatially distributed groundwater recharge rates in a tropical wet sedimentary region despite cloudy conditions

https://doi.org/10.1016/j.jhydrol.2022.127503Get rights and content

Highlights

  • A new satellite-based approach was proposed to estimate groundwater recharge rates.

  • Groundwater recharge was determined in a tropical wet region under cloudy condition.

  • Cloud-cleaning procedure enabled the production of a fair proxy for ET estimation.

  • Explicit representations of important root-zone water storage changes were shown.

  • Consistent validations showed the good performance of the satellite-based approach.

Abstract

Groundwater recharge (GWR) is one of the most challenging water fluxes to estimate, as it relies on observed data that are often limited in many developing countries. This study developed an innovative water budget method using satellite products for estimating the spatially distributed GWR at monthly and annual scales in tropical wet sedimentary regions despite cloudy conditions. The distinctive features proposed in this study include the capacity to address 1) evapotranspiration estimations in tropical wet regions frequently overlaid by substantial cloud cover; and 2) seasonal root-zone water storage estimations in sedimentary regions prone to monthly variations. The method also utilises satellite-based information of the precipitation and surface runoff. The GWR was estimated and validated for the hydrologically contrasting years 2016 and 2017 over a tropical wet sedimentary region located in North-eastern Brazil, which has substantial potential for groundwater abstraction. This study showed that applying a cloud-cleaning procedure based on monthly compositions of biophysical data enables the production of a reasonable proxy for evapotranspiration able to estimate groundwater by the water budget method. The resulting GWR rates were 219 (2016) and 302 (2017) mm yr−1, showing good correlations (CC = 0.68 to 0.83) and slight underestimations (PBIAS = −13 to −9%) when compared with the referenced estimates obtained by the water table fluctuation method for 23 monitoring wells. Sensitivity analysis shows that water storage changes account for +19% to −22% of our monthly evaluation. The satellite-based approach consistently demonstrated that the consideration of cloud-cleaned evapotranspiration and root-zone soil water storage changes are essential for a proper estimation of spatially distributed GWR in tropical wet sedimentary regions because of their weather seasonality and cloudy conditions.

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)

  • V. Lakshmi et al.

    A comparative study of available water in the major river basins of the world

    J. Hydrol.

    (2018)
  • M. Lal et al.

    Reverification of antecedent moisture condition dependent runoff curve number formulae using experimental data of Indian watersheds

    Catena

    (2019)
  • X. Lu et al.

    Correcting GPM IMERG precipitation data over the Tianshan Mountains in China

    J. Hydrol.

    (2019)
  • S.A. Mathias et al.

    Soil moisture data as a constraint for groundwater recharge estimation

    J. Hydrol.

    (2017)
  • M. Minnig et al.

    Impact of urbanization on groundwater recharge rates in Dübendorf, Switzerland

    J. Hydrol.

    (2018)
  • Q. Mu et al.

    Development of a global evapotranspiration algorithm based on MODIS and global meteorology data

    Remote Sens. Environ.

    (2007)
  • Q. Mu et al.

    Improvements to a MODIS global terrestrial evapotranspiration algorithm

    Remote Sens. Environ.

    (2011)
  • D.F. Rossetti et al.

    Late Quaternary sedimentation in the Paraíba Basin, Northeastern Brazil: landform, sea level and tectonics in Eastern South America passive margin

    Palaeogeogr. Palaeoclimatol. Palaeoecol.

    (2011)
  • F.A. Semire et al.

    Ground validation of space-borne satellite rainfall products in Malaysia

    Adv. Space Res.

    (2012)
  • Z.-H. Shi et al.

    Research on the SCS-CN initial abstraction ratio using rainfall-runoff event analysis in the Three Gorges Area, China

    CATENA

    (2009)
  • S. Veeck et al.

    Scale dynamics of the HIDROPIXEL high-resolution DEM-based distributed hydrologic modeling approach

    Environ. Modell. Software

    (2020)
  • E. Wendland et al.

    Water balance in the Guarani Aquifer outcrop zone based on hydrogeologic monitoring

    J. Hydrol.

    (2007)
  • W. Aeschbach-Hertig et al.

    Regional strategies for the accelerating global problem of groundwater depletion

    Nat. Geosci.

    (2012)
  • T. Ahmadi et al.

    Mapping groundwater recharge areas using CRD and RIB methods in the semi-arid Neishaboor Plain, Iran

    Arab. J. Geosci.

    (2015)
  • M. Ajmal et al.

    Quantifying Excess Stormwater Using SCS-CN–Based Rainfall Runoff Models and Different Curve Number Determination Methods

    J. Irrig. Drain. Eng.

    (2015)
  • R.G. Allen et al.

    Crop evapotranspiration - Guidelines for computing crop water requirements

  • Alley, W.M., Konikow, L.F., 2015. Bringing GRACE down to Earth. Groundwater 53, 826–829....
  • C.A. Alvares et al.

    Köppen’s climate classification map for Brazil

    Meteorol. Z.

    (2013)
  • Araújo Filho, J.C., Burgos, N., Lopes, O.F., Silva, F.H.B.B. Da, Medeiros, L.A.R., Melo Filho, H.F.R. De, Parahyba,...
  • R.S.A. Palharini et al.

    Climatological behavior of precipitating clouds in the northeast region of Brazil

    Adv. Meteorol.

    (2017)
  • A.G. Awadallah et al.

    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.

    (2017)
  • P. Brunner et al.

    How can remote sensing contribute in groundwater modeling?

    Hydrogeol. J.

    (2007)
  • T.J. Callahan et al.

    Estimating groundwater recharge in lowland watersheds

    Hydrol. Process.

    (2012)
  • CPRM – Geological Service of Brazil, 2020. SIAGAS – Groundwater Information System [WWW Document]. Law No....
  • R.S. Crosbie et al.

    Ground truthing groundwater-recharge estimates derived from remotely sensed evapotranspiration: a case in South Australia

    Hydrogeol. J.

    (2015)
  • R.S. Crosbie et al.

    Constraining the magnitude and uncertainty of specific yield for use in the water table fluctuation method of estimating recharge

    Water Resour. Res.

    (2019)
  • M.O. Cuthbert et al.

    Observed controls on resilience of groundwater to climate variability in sub-Saharan Africa

    Nature

    (2019)
  • P. Döll et al.

    Global-scale assessment of groundwater depletion and related groundwater abstractions: combining hydrological modeling with information from well observations and GRACE satellites

    Water Resour. Res.

    (2014)
  • P. Durán-Barroso et al.

    Sources of uncertainty in the NRCS CN model: recognition and solutions

    Hydrol. Process.

    (2017)
  • J.S. Famiglietti

    The global groundwater crisis

    Nat. Clim. Change

    (2014)
  • J.S. Famiglietti et al.

    Satellites provide the big picture

    Science

    (2015)
  • M. Gokmen et al.

    Assessing groundwater storage changes using remote sensing-based evapotranspiration and precipitation at a large semiarid basin scale

    J. Hydrometeorol.

    (2013)
  • I.E.M. Graaf et al.

    Advances in Water Resources A global-scale two-layer transient groundwater model: development and application to groundwater depletion

    Adv. Water Resour.

    (2017)
  • R.H. Hawkins et al.

    Runoff probability, storm depth, and curve numbers

    J. Irrig. Drainage Eng.

    (1985)
  • R.W. Healy et al.

    Using groundwater levels to estimate recharge

    Hydrogeol. J.

    (2002)
  • Huffman, G.J., Bolvin, D.T., Nelkin, E.J., 2018. Integrated Multi-satellitE Retrievals for GPM (IMERG) Technical...
  • S. Jasechko et al.

    The pronounced seasonality of global groundwater recharge

    Water Resour. Res.

    (2014)
  • M. Lal et al.

    Evaluation of the Soil Conservation Service curve number methodology using data from agricultural plots

    Hydrogeol. J.

    (2017)
  • H.M. Le et al.

    Comparison and bias correction of TMPA precipitation products over the Lower Part of Red-Thai Binh river basin of Vietnam

    Remote Sens.

    (2018)
  • G. Lenderink et al.

    Estimates of future discharges of the river Rhine using two scenario methodologies: direct versus delta approach

    Hydrol. Earth Syst. Sci.

    (2007)
  • Cited by (0)

    1

    These authors contributed equally to this work.

    View full text