Deep Learning-Based Forecasting of Groundwater Level Trends in India: Implications for Crop Production and Drinking Water Supply

Pragnaditya Malakar; Abhijit Mukherjee; Soumendra N. Bhanja; Sudeshna Sarkar; Dipankar Saha; Ranjan Kumar Ray

doi:10.1021/acsestengg.0c00238

Deep Learning-Based Forecasting of Groundwater Level Trends in India: Implications for Crop Production and Drinking Water Supply

Pragnaditya Malakar, Abhijit Mukherjee, Soumendra N. Bhanja, Sudeshna Sarkar, Dipankar Saha, Ranjan Kumar Ray

Research output: Contribution to journal › Article › peer-review

19 Scopus citations

Abstract

Despite numerous studies in recent times, there is no consensus on the primary drivers for groundwater storage (GWS) changes over India. Thus, predicting future groundwater level trends seems remote. In this context, using Gravity Recovery and Climate Experiment (GRACE)-derived GWS, WaterGap model-based groundwater recharge (GWR), and groundwater withdrawal (GWW), we show that GWW exhibits a stronger dominance than GWR on GWS change over India. Furthermore, we developed feed-forward neural network (FNN), recurrent neural network (RNN), and deep learning-based long short-term memory network (LSTM) models using multidepth in situ observations from a dense network of monitoring wells (n = 5367, 1996−2018), to simulate and forecast groundwater levels (GWL) in India. The result demonstrates the better performance of LSTM (>84% of observation wells showing r > 0.6, RMSE_n < 0.7) across India, outperforming both FNN and RNN in the testing period of 5 years (2014−2018). Our estimates also reveal that besides the prevailing long-term (1996−2018) statistically significant (p < 0.1) declining GWL trends in northwest India and the Ganges river basin, higher declining trends will potentially be observed in parts of north-central and south India in the forecasting period of 5 years (2019−2023). We envisage that the forecasting approach presented in the study can contribute toward an improved urban−rural drinking water supply and sustainable crop production for 1.3 billion people in India.(Figure

Original language	English
Pages (from-to)	965-977
Number of pages	13
Journal	ACS ES and T Engineering
Volume	1
Issue number	6
DOIs	https://doi.org/10.1021/acsestengg.0c00238
State	Published - Jun 11 2021
Externally published	Yes

Funding

Financial support for this work was provided by the Ministry of Human Resource Development, Government of India (project no: IIT/SRIC/GG & CSE/AGI/2013-14/201) and Department of Science and Technology, Government of India (project no: DST/TMD-EWO/WTI/2K19/EWFH/2019/ 201 (G) & (C) Dated: 28.10.2020). Notes The authors declare no competing financial interest. We acknowledge the freely available data from Central Ground Water Board (CGWB), Government of India; Climatic Research Unit, GRACE, Global Land Data Assimilation System (GLDAS), and WaterGAP. Groundwater level data are retrieved from the CGWB’s repositories, which could be accessed here: http://cgwb.gov.in/GW-data-access.html. Precipitation data were available from the archives of the Climatic Research Unit (CRU TS v-4.01) at https://crudata.uea.ac.uk/ cru/data/hg/. GRACE terrestrial water storage data were obtained from NASA JPL’s archive: http://grace.jpl.nasa.gov. The GLDAS data are obtained from GLDAS simulations that are archived and distributed by the Goddard Earth Sciences (GES) Data and Information Services Center (DISC) and can be accessed at https://disc.gsfc.nasa.gov/datasets?keywords= evapotranspiration&sort=-timeRes&page=1&source= Models%2FAnalyses%20OBSERVATION%20BASED. Groundwater withdrawals and groundwater recharge data are taken from the WaterGAP (version 2.2b) simulation, available from Goethe University Frankfurt’s archive (https://www.uni-frankfurt.de/45217892/Datens%C3%A4tze___Data_sets).

Funders	Funder number
Department of Science and Technology, Ministry of Science and Technology, India	DST/TMD-EWO/WTI/2K19/EWFH/2019/ 201, 28.10.2020
Ministry of Education, India	CSE/AGI/2013-14/201

Keywords

Groundwater quantity
India
LSTM-based forecasting
Relative driver importance on groundwater storage
water−food−energy nexus

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1021/acsestengg.0c00238

Cite this

@article{126090f3159746a08955b059064aa72b,

title = "Deep Learning-Based Forecasting of Groundwater Level Trends in India: Implications for Crop Production and Drinking Water Supply",

abstract = "Despite numerous studies in recent times, there is no consensus on the primary drivers for groundwater storage (GWS) changes over India. Thus, predicting future groundwater level trends seems remote. In this context, using Gravity Recovery and Climate Experiment (GRACE)-derived GWS, WaterGap model-based groundwater recharge (GWR), and groundwater withdrawal (GWW), we show that GWW exhibits a stronger dominance than GWR on GWS change over India. Furthermore, we developed feed-forward neural network (FNN), recurrent neural network (RNN), and deep learning-based long short-term memory network (LSTM) models using multidepth in situ observations from a dense network of monitoring wells (n = 5367, 1996−2018), to simulate and forecast groundwater levels (GWL) in India. The result demonstrates the better performance of LSTM (>84% of observation wells showing r > 0.6, RMSEn < 0.7) across India, outperforming both FNN and RNN in the testing period of 5 years (2014−2018). Our estimates also reveal that besides the prevailing long-term (1996−2018) statistically significant (p < 0.1) declining GWL trends in northwest India and the Ganges river basin, higher declining trends will potentially be observed in parts of north-central and south India in the forecasting period of 5 years (2019−2023). We envisage that the forecasting approach presented in the study can contribute toward an improved urban−rural drinking water supply and sustainable crop production for 1.3 billion people in India.(Figure",

keywords = "Groundwater quantity, India, LSTM-based forecasting, Relative driver importance on groundwater storage, water−food−energy nexus",

author = "Pragnaditya Malakar and Abhijit Mukherjee and Bhanja, {Soumendra N.} and Sudeshna Sarkar and Dipankar Saha and Ray, {Ranjan Kumar}",

note = "Publisher Copyright: {\textcopyright} 2021 American Chemical Society.",

year = "2021",

month = jun,

day = "11",

doi = "10.1021/acsestengg.0c00238",

language = "English",

volume = "1",

pages = "965--977",

journal = "ACS ES and T Engineering",

issn = "2690-0645",

publisher = "American Chemical Society",

number = "6",

}

TY - JOUR

T1 - Deep Learning-Based Forecasting of Groundwater Level Trends in India

T2 - Implications for Crop Production and Drinking Water Supply

AU - Malakar, Pragnaditya

AU - Mukherjee, Abhijit

AU - Bhanja, Soumendra N.

AU - Sarkar, Sudeshna

AU - Saha, Dipankar

AU - Ray, Ranjan Kumar

PY - 2021/6/11

Y1 - 2021/6/11

N2 - Despite numerous studies in recent times, there is no consensus on the primary drivers for groundwater storage (GWS) changes over India. Thus, predicting future groundwater level trends seems remote. In this context, using Gravity Recovery and Climate Experiment (GRACE)-derived GWS, WaterGap model-based groundwater recharge (GWR), and groundwater withdrawal (GWW), we show that GWW exhibits a stronger dominance than GWR on GWS change over India. Furthermore, we developed feed-forward neural network (FNN), recurrent neural network (RNN), and deep learning-based long short-term memory network (LSTM) models using multidepth in situ observations from a dense network of monitoring wells (n = 5367, 1996−2018), to simulate and forecast groundwater levels (GWL) in India. The result demonstrates the better performance of LSTM (>84% of observation wells showing r > 0.6, RMSEn < 0.7) across India, outperforming both FNN and RNN in the testing period of 5 years (2014−2018). Our estimates also reveal that besides the prevailing long-term (1996−2018) statistically significant (p < 0.1) declining GWL trends in northwest India and the Ganges river basin, higher declining trends will potentially be observed in parts of north-central and south India in the forecasting period of 5 years (2019−2023). We envisage that the forecasting approach presented in the study can contribute toward an improved urban−rural drinking water supply and sustainable crop production for 1.3 billion people in India.(Figure

AB - Despite numerous studies in recent times, there is no consensus on the primary drivers for groundwater storage (GWS) changes over India. Thus, predicting future groundwater level trends seems remote. In this context, using Gravity Recovery and Climate Experiment (GRACE)-derived GWS, WaterGap model-based groundwater recharge (GWR), and groundwater withdrawal (GWW), we show that GWW exhibits a stronger dominance than GWR on GWS change over India. Furthermore, we developed feed-forward neural network (FNN), recurrent neural network (RNN), and deep learning-based long short-term memory network (LSTM) models using multidepth in situ observations from a dense network of monitoring wells (n = 5367, 1996−2018), to simulate and forecast groundwater levels (GWL) in India. The result demonstrates the better performance of LSTM (>84% of observation wells showing r > 0.6, RMSEn < 0.7) across India, outperforming both FNN and RNN in the testing period of 5 years (2014−2018). Our estimates also reveal that besides the prevailing long-term (1996−2018) statistically significant (p < 0.1) declining GWL trends in northwest India and the Ganges river basin, higher declining trends will potentially be observed in parts of north-central and south India in the forecasting period of 5 years (2019−2023). We envisage that the forecasting approach presented in the study can contribute toward an improved urban−rural drinking water supply and sustainable crop production for 1.3 billion people in India.(Figure

KW - Groundwater quantity

KW - India

KW - LSTM-based forecasting

KW - Relative driver importance on groundwater storage

KW - water−food−energy nexus

UR - http://www.scopus.com/inward/record.url?scp=85117060292&partnerID=8YFLogxK

U2 - 10.1021/acsestengg.0c00238

DO - 10.1021/acsestengg.0c00238

M3 - Article

AN - SCOPUS:85117060292

SN - 2690-0645

VL - 1

SP - 965

EP - 977

JO - ACS ES and T Engineering

JF - ACS ES and T Engineering

IS - 6

ER -

Deep Learning-Based Forecasting of Groundwater Level Trends in India: Implications for Crop Production and Drinking Water Supply

Abstract

Funding

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this