Forecasting of Groundwater Level Variation Under Changing Climate in Chhattisgarh State Using Deep Learning Technique

Forecasting of Groundwater Level Variation Under Changing Climate in Chhattisgarh State Using Deep Learning Technique | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Forecasting of Groundwater Level Variation Under Changing Climate in Chhattisgarh State Using Deep Learning Technique Mukesh Kumar Dey, Chandan Kumar Singh This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3927808/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Groundwater is the most precious natural resource in modern days. India is the largest consumer of groundwater globally, with over 25% of the world's groundwater extraction. Climate change affects the groundwater level both in direct and indirect ways. Recently developed machine learning approaches have led to the consideration of selected climate variables that can govern the groundwater dynamic. The inclusion of indirect key drivers such as anthropogenic activities and lithology to forecast groundwater levels using machine learning techniques is not well understood. This paper aims to consider both the direct and indirect key drivers for forecasting seasonal groundwater levels. In this context, a modified approach based on a deep learning model has been formulated that considers land cover dynamics, lithological properties, and climatic variables such as temperature and precipitation. The model was calibrated and validated to forecast seasonal groundwater levels for four Shared Socioeconomic Pathways (SSPs) scenarios. The results show that the accuracy level, R 2 is 0.86 which is acceptable. Overall, the results obtained broadly correspond to an acceptable degree of accuracy. The proposed methodology is applicable for seasonal groundwater level forecasting and can be useful to farmers and key stakeholders. Climate change Groundwater Forecasting Deep learning Neural network CMIP6 Land Use Land Cover Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 Figure 14 Figure 15 Figure 16 Figure 17 1. Introduction Climate change is the long-term changes of climatic factors. In the 21st century, humans are the most responsible ones behind climate change (Nations 2022). Climate change's impact on natural resources is quite notable. The urgency of the global climate change problem has grown so, investigating climatic conditions and variables has become increasingly popular. The problem of climate change in the world becomes even more critical, so, there has been a growing interest in investigating climate characteristics and factors. Groundwater is one of the most precious natural resources. In India, human activities rely more on groundwater than surface water. Agriculture sectors in India directly depend on groundwater and the ratio of use groundwater and surface water is increasing year by year (NSSO 2014). The vast use of groundwater results in fluctuations in groundwater levels in pre-monsoon and post-monsoon. These abrupt fluctuations in groundwater affect the farmers to estimate their area sowing crops. The precise climate data that can be used to create efficient strategies for both mitigation and adaptation is the greatest method for communities to adjust to changing climatic circumstances (Reddy and Saravanan 2023). Therefore, exact and trustworthy groundwater level (GWL) predictions, offer crucial data on the future availability of groundwater and may be used as the foundation for management choices and plans. Climate models are constantly being updated. Models updated in terms of higher spatial resolution, new physical processes, and biogeochemical cycles (Bader et al. 2008). In 2021, the IPCC released its sixth assessment report, which included CMIP6 climate models. In contrast to its predecessors, the most recent CMIP6 GCMs produce a considerably more accurate picture of Earth's physical processes (Calvin et al. 2023). The CMIP6 models in particular use SSPs to forecast future occurrences (Schlund et al. 2020). Climatic factors such as Precipitation and Temperature vary in long-term future different climatic scenarios. Shared Socioeconomic Pathways (SSPs) scenarios are created by an international team of climate scientists and economists to give the climate change research community guidance for doing an integrated, multidisciplinary examination (Riahi et al. 2017). According to the IPCC study, the SSPs have five storylines, including sustainable development, regional competition, inequality, fossil-fueled development, and middle-of-the-road development (UNEP 2022). CMIP6 model comes with a certain region-wise bias (Yeboah et al. 2022). We used biased corrected data which is freely available and open for use (Mishra et al. 2020). 13 CMIP6 models are available for India. Each model comes with four SSP scenarios (SSP126, SSP245, SSP370, SSP585) and historical data. This data comes with 0.25 0 * 0.25 0 spatial resolution and three parameters i.e., precipitation, t max , t min . The artificial neural network (ANN) approach has recently proved its value in forecasting groundwater levels solely based on meteorological variables (Joshi et al. 2020). Recent studies have concentrated on the viability of using machine learning (ML) methods, particularly the branch of ML known as deep learning (DL), in different subfields of hydrology. Some ultramodern engineering uses of ML, such as RL and other innovative methods for applying ML to conventional prediction problems, have not yet been explored. The best approaches to use this geospatial temporal dataset and enhance integrated hydrological system modeling are provided by technologies that belong under the category of geospatial and geo-spatiotemporal AI, such as DL and parallel computing (Ghobadi and Kang 2023). First off, earlier research concentrated on traditional ML models, but the more recently developing DL attention-based models (such as long- and short-term time-series networks (LSTNet), transformer, informer, and conformer) are still in their infancy, especially in the field of WRM (Agarap 2019). CGWB is the main central apex agency in India for groundwater. CGWB takes seasonal readings of groundwater levels all over India. CGWB has various observation wells in India to take reading seasonally. In the generation of Remote Sensing and GIS, CGWB takes groundwater level readings manually. Chhattisgarh states have 1098 observation wells. CGWB dept. Under govt. of India takes groundwater level readings in these wells in 4 seasons in a year- monsoon (June-September), Post-monsoon (October-November), Winter (December-February), and Pre-monsoon (March-May). GW level measurement manually of observation wells at a large scale in Chhattisgarh is very time-consuming and not cost-beneficial at all. In this study, we investigated both the direct and indirect key drivers for forecasting seasonal groundwater levels. To the author’s knowledge, the inclusion of indirect drivers such as anthropogenic activities and lithology to forecast groundwater levels using machine learning techniques has been scarcely investigated (Obahoundje et al. 2017). To address this issue, the present study explores both the direct and indirect drivers for forecasting seasonal groundwater levels in the Chhattisgarh state, India. We designed a deep learning-based model that considers land cover dynamics, lithological properties, and climatic variables such as temperature and precipitation to forecast seasonal groundwater model. We use the ANN approach to predict seasonal groundwater levels with high accuracy using precipitation, temperature, land use land cover percentage, and lithology condition as inputs. The calibration and validation of the model were performed using historical climate data. We forecasted and compared the change in groundwater levels for four Shared Socioeconomic Pathways (SSPs) scenarios. This will give a proper idea of how the groundwater level will be in upcoming years and how we can efficiently use it in different sections to maintain the sustainability of groundwater use. This paper presents an innovative method that can be applied to farmers and key stakeholders for effective planning and management of groundwater. 2. Data and study area 2.1 Study Area Description : - The area of interest in this study is Chhattisgarh state. Chhattisgarh, state of east-central India. It is bounded by the Indian states of Uttar Pradesh and Jharkhand to the north and northeast, Odisha (Orissa) to the east, Telangana (formerly part of Andhra Pradesh) to the south, and Maharashtra and Madhya Pradesh to the west (Fig. 1 ). Its capital is Raipur. Area 52,199 square miles (135,194 square km). Pop. (2011) 25,540,196. The rectangular extent of Chhattisgarh state is 80.24 E, 17.78 N, 84.39 E, 24.10 N. Chhattisgarh state has five major divisions i.e., Surguja division, Bilaspur division, Durg division, Raipur division, Bastar division. Each division has a specific district. According to the latest report, a total of 33 districts are presently available in Chhattisgarh. The central part of Chhattisgarh is our area of interest which mainly consists of 14 major districts. The major districts are- Bastar, Bilaspur, Dhamtari, Durg, Janjgir-Champa, Jashpur, Kanker, Kawardha, Korba, Koriya, Mahasamund, Raigarh, Raipur and Rajnandgaon. 2.2 Data Selection : - The required data were collected from different websites and the CGWB office which is govt. body of India as shown in Table 1 . Table 1 Data used in this study Sr. No. Data Components Resolutions Sources 1 IMDAA data Precipitation (12 km*12 km) https://rds.ncmrwf.gov.in/datasets 2 m Temperature 2 Bias-corrected CMIP6 data Precipitation (0.25 0 * 0.25 0 ) https://zenodo.org/record/3873998#.Y1pQtnZBzIX Max. Temperature Min. Temperature 3 Groundwater level field data Well depth Observation well wise CGWB office, Raipur Water level 4 LULC Land Use Land Cover Change in Chhattisgarh 30 m LANDSAT Satellite 5 Lithology Condition Lithology type Observation well wise CGWB office, Raipur Groundwater level (GWL) data from 1995 − 2020 for 267 observation wells have been used in this study. The maximum depth of the observation wells is 290 m. The well type varies from dug wells and bore wells to tube wells. Central Ground Water Board (CGWB), the main body of India measures groundwater levels all over India. They take groundwater level readings in these wells in 4 seasons in a year- monsoon (June-September), Post-monsoon (October-November), Winter (December-February), and Pre-monsoon (March-May). We used two climatic variables i.e., Precipitation and Temperature with spatial resolution of 12 Km × 12 Km for historical data and 0.25 0 × 0.25 0 (∼27.5 × 27.5 km in the equation) spatial resolution for future climate data. The future climatic data is biased corrected and downscaled, and it is freely available (Mishra et al. 2020 ). The lithology describes the physical characteristics of the rocks. The central part of Chhattisgarh has 9 types of lithology conditions (Fig. 2 ). The types are- basalt, charnokite, gnesis, gneissic Complex- basement, granite, limestone, sandstone, schist, and shale. Each lithology has its properties and different hydraulic conductivity values. We used the hydraulic conductivity value of the lithology types. Lithology is the physical characteristics of rocks and most of them are impermeable, so the water-holding capacity is very low. That’s why the hydraulic conductivity value of rocks is very low which is mentioned below (Table 2 ). Table 2 Hydraulic conductivity value of each Lithology (Widodo et al. 2016 ) Sr. No. Types of Lithology Hydraulic conductivity(k) m/s I. Gneissic Complex-Basement 0.000002164 II. Shale 0.000000411499 III. Limestone 0.00000163791 IV. Gneiss 0.000002164 V. Sandstone 0.00000190922 VI. Basalt 0.0001018519 VII. Charnokite 0.00005956 VIII. Granite 0.0000041 IX. Schist 0.000000694444 Land use Land cover outlines the spatial information of distinct categories of the cover over the earth's surface. It identifies how the area is being used and detects the type of cover. Several types of LULC can be water bodies, bare soil, urban areas, agricultural areas, forest cover, etc. This input data plays a significant role in the amount of flooding in the area (Oxoli et al. 2018 ). In the current study land use land cover was done by supervised classification. The training data used for supervised classification was the ESRI LULC map for 2020. It was downloaded from the ESRI website. Thirty-year Land Use Land cover was prepared in the interval of 5 years i.e. from 2000 to 2020. Figure 3 shows the land use land cover changes over Chhattisgarh state from year 2000 to 2020. The LULC map is made in a systematic manner using the Colab environment. Raw data for the classification of the area was LANDSAT Satellite imageries (Manandhar et al. 2009 ). Making use LANDSAT 9 LULC for the year 2020 was prepared, while LANDSAT 8 was used for the year 2015, and for the years 2010, 2005, 2000, and 1995 LANDSAT 5 imageries were used. The Land use land cover was prepared in the Colab environment using GEE (Google Earth engine). Colab is a cloud-based Jupyter notebook that runs in the personal Google workbook. It can access the web and the internet. Python code can be written and executed in such an environment. A geospatial processing service is called Google Earth Engine. With the help of Earth Engine and the Google Cloud Platform, you can carry out geospatial processing at a scale. Earth Engine is designed to create a platform that is interactive for large-scale geospatial algorithm development, making significant advancements in addressing global concerns using big geospatial datasets. Making use of these two tools land use land cover was prepared and the area was classified into seven classes namely water, cropland, bare land, forest area, built-up area, shrubs, and other vegetation. 3. Materials and Methodology 3.1 Methodology : - Figure 4 outlines the modified approach to forecasting seasonal groundwater levels using a deep learning-based technique. Contrary to the conventional approach that makes use of limited input, we introduced a modified approach that combines both climate data and indirect drivers. The rainfall and temperature are the main climatic factors that drive the model. The output is based on climatic variables. The input variables are rainfall and temperature as a time series. Based on the input variable the model is calibrated and then after the validation of calibration, it predicts the output variable. The modified approach, which is the main method of this study, is based on a combination of climatic factors, lithology conditions land use, and land cover changes over the years. We took five variables as our input variables. The input variables are- rainfall, temperature, hydraulic conductivity of the lithology condition, and the land use land cover. The input variables are converted into usable format from their raw format using Python in the pre-processed step. The input variables are then sorted as a time-series pattern to supply the model. Then we separate the input variable into two parts- features and levels. The feature is the individual, independent variables serving as. To create predictions, prediction models require characteristics. A feature engineering approach may also be used to create new features from existing ones. The label is the finished product. The output classes refer to the labels. Data scientists refer to a collection of samples that have been assigned to one or more labels when they talk about labeled data. The time series is now Split into two parts- the training set and the test set. A piece of our real dataset called the "training set" is fed into the deep learning model to help it find and understand patterns. This develops the model. Usually, training data is bigger than testing data. This is because we want to provide the model with as much information as we can for it to identify and learn useful patterns. When our dataset’s data are supplied to a machine learning algorithm, this recognizes patterns in the data and concludes. The test set analyses the effectiveness and development of your training algorithms and alters or optimizes them for better outcomes. Two primary standards apply when testing data. It should be a true representation of the dataset and be big enough to produce accurate predictions. Test data offer a last-minute, practical verification of an unknown dataset to show that the deep learning algorithm was successfully trained. 3.2 Model development and performance : - The proposed approach is applied as a filter to increase the efficacy of hydrological models. Almost all scientific and technical fields have effectively used deep learning as a tool (Rozos et al. 2022 ). Machine learning models have advanced in hydrology from basic feed-forward networks used for short-term forecasting to complicated models that can consider even the static properties of catchments, mimicking the hydrological experience (Ghobadi and Kang 2023 ). According to recent studies, machine learning models are more reliable and effective than conventional hydrological models (both conceptual and physically based), often exceeding them (Ghobadi and Kang 2023 ). An indirect approach to take advantage of ML in hydrological applications is to use it as a tool for pre-processing the data or post-processing the results of the standard hydrological models (Panahi et al. 2020 ). To build these types of models we first need to understand the nature of data. In our dataset, there are 2 types of data are there- features and labels. The label data is the most important data to calibrate and validate our results. To calibrate the model, it’s particularly important to set the hyperparameter accurately. The hyper-parameters were adapted based on trial-and-error methods. We used various hyper-parameters to ensure that the model works accurately. First, we choose the activation in this model as an ‘Adam’ optimizer with a learning rate of 0.01. The batch size and epoch size are 33 and 2000, respectively. We chose the activation function as a Rectified linear unit (relu) which depends upon the best fit curve of the groundwater level in each well. Figure 5 Shows the best fit curve is lognormal. Similarly, we check it for all the observation wells. The largest wells show the results in between lognormal and gamma. According to this distribution function, we choose our activation function as ‘relu’. Although some wells show normal distribution. The total GWL data set is taken for 1995–2020 i.e., 26 years (26 × 4 = 104 seasons). This dataset is split into two parts: the training set is taken 15 years i.e., from 1995 to 2009 and the testing data is set for the remaining years from 2010 to 2020, 11 years. Lastly, we forecasted the GWL up to the end of the 21st century (2100). The calibrated and validated results are checked by Nash-Sutcliffe Efficiency (NSE) and squared Pearson’s correlation coefficient (R 2 ) matrix. After the usable NSE and R 2 value we used it for future forecasting. Good NSE and R 2 value means the actual value of groundwater and the forecasted value is nearly the same for these wells. 4. Results and Discussions We applied our model to the future data. We sorted our future data according to the observation well and different SSP conditions. We all know that groundwater is a highly fluctuating earth property. As the climate changes very rapidly, the trend of groundwater level of also changing. Figure 6 shows that the annual trend of precipitation will change drastically in the future among all the SSPs. The blue line in the Fig. 6 is the historical precipitation in the Chhattisgarh state. The red dotted line is the worst SSP condition, and we also observed it in the below fig. The maximum change in precipitation is shown in this SSP condition. The Figs. 7 and 8 shows the annual tread of max. and min. is over the Chhattisgarh state. The max. temperature is above 37 0 C at the end of the 21st century. The min. temperature recorded at 27 0 C at the end of the 21st century. All the SSPs in fig. show increasing order. Calibration and validation performance We take 60 percent of our input data as training data and the remaining 40 percent of our data as test data. We also computed the actual and model forecasted (2010 − 2020) GWL time series with future forecasting up to the end of the 21st century (2021 − 2100) for all the wells. Figure 9 , Fig. 10 , and Fig. 11 summarizes the validation performance of our model. The below figures not only match the shapes but the peak points also capture very well. These wells show an accuracy level above 0.80 which is a good one. SSP wise forecasting Our goal is to forecast the future groundwater level in different SSP scenarios. In our study, we take 4 SSP scenarios i.e., SSP 126, SSP 245, SSP 370, and the worst condition SSP 585. The different SSP conditions have different carbon emissions, depending upon it has different forecasted precipitation and temperature values. We apply our calibrated model in these SSP scenarios. We want to visualize the change in groundwater level in different future scenarios. This will give a proper idea of how the groundwater level will be in upcoming years and how we can efficiently use it in different sections to maintain the sustainability of groundwater use. SSP 126 The SSP 126 is one of the good future scenarios. According to the SSP126 scenario, a sustainable future society will develop with significant economic growth, investments in health and education, and strong governance (Siabi et al. 2023 ). This scenario anticipates a 1.8 C increase in global average temperature by 2100, which is in line with the Paris Agreement's goals of limiting warming to below 2.0 C (Rogelj et al. 2016 ). Figure 12 shows the forecasted groundwater level of Chhattisgarh state of two different wells according to SSP 126. Fig shows the different season-wise groundwater levels in different colors. The green line which is season 3 (June- August) which is monsoon season where rainfall is maximum shows groundwater level minimum among all the seasons. Season 1 (December - January) and season 4 (September- November) groundwater levels in the range of 5 to 8 m. Sometimes the middle of the 21st century and season 4 touched a groundwater level of 10 m which is quite exceptional. The maximum level of groundwater forecasted by the model is in season 2 (February - May) which is the summer season where the precipitation is minimal. The forecasted groundwater level is in the range of 9 m to 11 m. The mean value of this season is around 10 m. The orange line of the above plot does not vary abruptly apart from 2060 to 2068. Here the groundwater level is around 9 m. The SSP 126 is the optimistic evolution of future society (O’Neill et al. 2016 ). So, the groundwater level is also in a certain range which is in the controllable situation. SSP 245 The SSP 245 scenario is called the “ middle of the road ” scenario (Fricko et al. 2017 ). This scenario is consistent with the climate initiatives taken by different nations to uphold their current obligations to cut emissions. The total radiative forcing is anticipated to increase in SSP 245 to around 4.5 Wm − 2 before declining after 2070 (Siabi et al. 2023 ). This leads to increased heat extremes and catastrophic flooding from precipitation extremes also predicted to cause a drop in food production throughout the planet. Figure 13 shows the forecasted groundwater level according to SSP 245. Here the groundwater level changing treads are somehow different from SSP 126 (IPCC 2000; harrisson 2018). The green color line which is season 3 shows groundwater level in the range of 2 to 6 m. This is the minimum level among all the other SSP scenarios. In this season the groundwater level fluctuates most. This is the monsoon season where the precipitation is also not uniform, this can be one of the reasons for this fluctuation. The summer seasons show the level in the range of 9 m to 10 m. In this season, in the last couple of years of the 21st century, the level slightly increases. Those years show the groundwater level above 10 m. The mean groundwater level of this season in SSP 245 is around 10 m. As this SSP is in the middle of all the SSP scenarios, the groundwater level also not touching extremely high in this SSP scenario. SSP 370 The SSP 370 scenario is the 3rd most common scenario where the burning of coal we are increasing slowly. Countries have a propensity to withdraw from international collaboration and concentrate on their economic aims in the SSP 370 scenario. Methane, aerosols, and other gases that trap heat other than carbon dioxide are anticipated to reach high concentrations (Meinshausen et al. 2011a ). By 2100. According to this scenario, there will be 12.6 billion people on the planet. Radiative forcing levels are predicted to increase to roughly 7.0 Wm − 2 at the end of 2100. Sea levels will rise drastically in this SSP. According to this SSP, the sea level will be between 46 and 74 cm (Riahi et al. 2017 ). Figure 14 shows the forecasted groundwater level according to SSP 370. The groundwater levels in both season 1 and season 4 overlapped with each other like the previous two SSP forecasting. The range of fluctuating groundwater water levels is between 6 m to 8 m. In the year between 2066 to 2069, the groundwater level suddenly decreased to 4 m for both observations. From Fig. 6 we can observe that in those years the precipitation drastically increased. Figures 7 and 8 show that the trends of maximum and minimum temperature are also in decreasing order. That is why the groundwater level also decreased. The groundwater level in the summer season is in the range of 8 m to 10 m. But at the end of 2100 years, the groundwater level shows slowly increasing order. From the year 2086 to 2100 the groundwater level shows an increase in order. At the end of the 21st century, it touches almost 11 m. SSP 585 In the SSP 585 scenario, humanity generally does nothing to stop climate change but continues to worsen it. In this scenario, burning coal, natural gas, and oil will all contribute to global economic growth. Radiative forcing levels are predicted to increase to roughly 8.5 Wm − 2 in this scenario (Meinshausen et al. 2011b ). Large-scale coastal flooding and incredibly damaging storms are expected in this scenario. Parts of the world will become uninhabitable as a result, particularly during the warmest times of the year. This will be the worst scenario. The average temperature rise will be 3.7 0 C in this SSP. According to the IPCC report, 2 0 C temperature rise is recognized as the threshold at which climate change becomes dangerous. The sea level rise will be 0.63 m. These effects will lead to the adaptation cost. In this SSP scenario, the adaptation cost will be much higher than other SSP scenarios. Figure 15 shows the forecasted groundwater level according to SSP 585. In this SSP scenario, all the seasons show increasing trends. The groundwater level in monsoon seasons is also quite high and increasing trends. In the monsoon season, the level is in the range of 2 m to 8 m. The groundwater level fluctuates a lot in this season. In season 1 (December- January) the groundwater level touched around 10 m which is the maximum in all the SSPs. In this SSP the climatic variable also shows an increase throughout the period, that’s the reason the groundwater level also increased. The summer season shows the highest groundwater level. The average groundwater level will be 11 m in this SSP scenario. In season 2 the groundwater level touched almost 12 m at the end of 2100. This will lead to groundwater scarcity in the upcoming years. Baseline comparison The season-wise comparison for each SSP with a baseline line is shown in Fig. 16 . At The baseline we take the last observed groundwater level. The last observed groundwater level was taken in the year 2020 by CGWB. We compare the groundwater level from the line with each SSP scenario and we can visualize the changes. The historical groundwater maximum in season 2 (February -May). Apart from SSP 585, all the other SSPs i.e., 126, 245, and 370 show the forecasted groundwater level is maximum in Season 2 itself. In SSP 585 season 1 shows the maximum ground water level. The orange bar in the plot denotes the season 3 groundwater level. It is the minimum in baseline and all the SSPs scenarios. The peaks are much higher in SSP 585 compared to all other SSPs. This will be more climate change-affected scenarios among all the others. In this SSP the precipitation pattern also drastically changes with max. and min. temperature. This affects the groundwater level that we can observe from the above plot. Percentage change with baseline Figure 17 shows the percentage change of groundwater level in different SSP scenarios season-wise. These changes we compare with the Baseline mean groundwater level in the year 2020. We can see the changes in the Fig. 17 . The groundwater levels are drastically changing over the year. The maximum changes are noted in SSP 585 and SSP 245. In these two scenarios, the groundwater level changes by more than 350%. In SSP 585 all the seasons show groundwater increasing trends of more than a minimum 150%. The season4 shows an increasing trend in all the SSPs. In this season the precipitation and temperature show increasing in order and the pattern of precipitation also changes. These changes a visualized in Fig. 6 . These changes cause more Vulnerability in the upcoming future years. Conclusions The prime objective of this study was to introduce a modified approach using a machine learning technique that considers the climate variables along with additional key factors. This study has focused on assessing the impacts of climate-changing scenario and their effects on groundwater levels. In this study, we forecasted groundwater levels using ANN models. This deep learning-based model forecasts accurately in those observation wells where the historical data is available and minimum data gaps are there. The data gaps also play a significant role in predicting the groundwater level. This study proves that the aquifer condition land use and land cover change also play and key role in the fluctuation of the groundwater level other than the rainfall and temperature. The more permeable aquifer condition shows satisfactory results other than rocks that are impermeable or semi-permeable. The changes in land use and land cover are also going to be responsible for groundwater level decline in the future. The increase of built areas and decreasing order of vegetation also affect the groundwater level. This affects the recharge area which can also handle groundwater level changes. Different SSP conditions show different groundwater level values. Although it is confirmed that the groundwater level for the upcoming future scenario will decline year by year. Each year the season 3 period means the monsoon period (June to August) shows groundwater in the range of 3 to 6 m. The summer season (February to May) shows groundwater in the range of 8 to 12 m. The other two seasons - season 1 (December to January) and season 4 (September to November) show groundwater levels in the range of 6 to 9 m. One of the main important observations of our study is that- the more permeable lithology condition of the observation wells shows better results than impermeable or semi-permeable aquifer-type observation wells. This study may in the future extend to climate change impacts on extreme events such as floods and droughts in terms of their frequency and occurrence should be analyzed. In this study we only used one model which was selected by literature review and earlier works have been done in central India, based on that we selected this model. As we forecast groundwater levels seasonally. This forecasting will help the farmers to irrigate their land according to the available groundwater in these regions. This forecast will be very beneficial for farmers, and key stakeholders such as the state water resources department, Central Groundwater Board (CGWB), etc. to maintain the recharge-discharge ratio by planning annual groundwater discharge for sustainable development. The farmers may get an appropriate information idea for irrigating their land based on the seasonal availability of groundwater for irrigation purposes. This enhanced information about groundwater level forecasts may help in short-term and long-term planning and management of the groundwater. Declarations Author Contributions: Both authors contributed to the study at all levels and original draft preparation. Mukesh Kumar Dey : Methodology, Visualization, Writing – original draft; Chandan Kumar Singh : Conceptualization, Writing – review, editing final draft. Funding: No funding was used in this research. Availability of Data and Materials: Data will be available upon reasonable request. Declarations Competing interests: The authors declare no competing interests. References Agarap AF (2019) Deep Learning using Rectified Linear Units (ReLU) Bader D, Covey C, Gutowski W, et al (2008) Climate Models: An Assessment of Strengths and Limitations. Climate Models: An Assessment of Strengths and Limitations Calvin K, Dasgupta D, Krinner G, et al (2023) IPCC, 2023: Climate Change 2023: Synthesis Report. Contribution of Working Groups I, II and III to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change [Core Writing Team, H. Lee and J. Romero (eds.)]. IPCC, Geneva, Switzerland. Intergovernmental Panel on Climate Change (IPCC) Fricko O, Havlik P, Rogelj J, et al (2017) The marker quantification of the Shared Socioeconomic Pathway 2: A middle-of-the-road scenario for the 21st century. Global Environmental Change 42:251–267. https://doi.org/10.1016/j.gloenvcha.2016.06.004 Ghobadi F, Kang D (2023) Application of Machine Learning in Water Resources Management: A Systematic Literature Review. Water 15:620. https://doi.org/10.3390/w15040620 harrisson thomas (2018) Explainer: How ‘Shared Socioeconomic Pathways’ explore future climate change. In: Carbon Brief. https://www.carbonbrief.org/explainer-how-shared-socioeconomic-pathways-explore-future-climate-change/. Accessed 13 Jan 2024 IPCC (ed) (2000) Emissions scenarios: summary for policymakers;a special report of IPCC Working Group III$Intergovernmental Panel on Climate Change. Intergovernmental Panel on Climate Change Joshi N, Rahaman M, Thakur B, et al (2020) Assessing the Effects of Climate Variability on Groundwater in Northern India. World Environmental and Water Resources Congress Manandhar R, Odeh IOA, Ancev T (2009) Improving the Accuracy of Land Use and Land Cover Classification of Landsat Data Using Post-Classification Enhancement. Remote Sensing 1:330–344. https://doi.org/10.3390/rs1030330 Meinshausen M, Smith SJ, Calvin K, et al (2011a) The RCP greenhouse gas concentrations and their extensions from 1765 to 2300. Climatic Change 109:213–241. https://doi.org/10.1007/s10584-011-0156-z Meinshausen M, Smith SJ, Calvin K, et al (2011b) The RCP greenhouse gas concentrations and their extensions from 1765 to 2300. Climatic Change 109:213. https://doi.org/10.1007/s10584-011-0156-z Mishra V, Bhatia U, Tiwari AD (2020) Bias-corrected climate projections for South Asia from Coupled Model Intercomparison Project-6. Sci Data 7:338. https://doi.org/10.1038/s41597-020-00681-1 Nations U (2022) What Is Climate Change? In: United Nations. https://www.un.org/en/climatechange/what-is-climate-change. Accessed 10 Mar 2023 NSSO (2014) Statistics on Indian Economy and Society. https://www.indianstatistics.org/irrigation.html. Accessed 4 Jan 2023 Obahoundje S, Ofosu EA, Akpoti K, Kabo-bah AT (2017) Land Use and Land Cover Changes under Climate Uncertainty: Modelling the Impacts on Hydropower Production in Western Africa. Hydrology 4:2. https://doi.org/10.3390/hydrology4010002 O’Neill BC, Tebaldi C, Van Vuuren DP, et al (2016) The Scenario Model Intercomparison Project (ScenarioMIP) for CMIP6. Geosci Model Dev 9:3461–3482. https://doi.org/10.5194/gmd-9-3461-2016 Oxoli D, Ronchetti G, Minghini M, et al (2018) Measuring Urban Land Cover Influence on Air Temperature through Multiple Geo-Data—The Case of Milan, Italy. ISPRS International Journal of Geo-Information 7:421. https://doi.org/10.3390/ijgi7110421 Panahi M, Sadhasivam N, Pourghasemi HR, et al (2020) Spatial prediction of groundwater potential mapping based on convolutional neural network (CNN) and support vector regression (SVR). Journal of Hydrology 588:125033. https://doi.org/10.1016/j.jhydrol.2020.125033 Reddy NM, Saravanan S (2023) Extreme precipitation indices over India using CMIP6: a special emphasis on the SSP585 scenario. Environ Sci Pollut Res. https://doi.org/10.1007/s11356-023-25649-7 Riahi K, van Vuuren DP, Kriegler E, et al (2017) The Shared Socioeconomic Pathways and their energy, land use, and greenhouse gas emissions implications: An overview. Global Environmental Change 42:153–168. https://doi.org/10.1016/j.gloenvcha.2016.05.009 Rogelj J, Den Elzen M, Höhne N, et al (2016) Paris Agreement climate proposals need a boost to keep warming well below 2 °C. Nature 534:631–639. https://doi.org/10.1038/nature18307 Rozos E, Dimitriadis P, Bellos V (2022) Machine Learning in Assessing the Performance of Hydrological Models. Hydrology 9:5. https://doi.org/10.3390/hydrology9010005 Schlund M, Lauer A, Gentine P, et al (2020) Emergent constraints on equilibrium climate sensitivity in CMIP5: do they hold for CMIP6? Earth Syst Dynam 11:1233–1258. https://doi.org/10.5194/esd-11-1233-2020 Siabi EK, Awafo EA, Kabo-bah AT, et al (2023) Assessment of Shared Socioeconomic Pathway (SSP) climate scenarios and its impacts on the Greater Accra region. Urban Climate 49:101432. https://doi.org/10.1016/j.uclim.2023.101432 UNEP W (2022) IPCC report: Climate Change 2022 Widodo L, Cahyadi T, Notosiswoyo S, Widijanto E (2016) Application of Clustering System to Analyze Geological, Geotechnical and Hydrogeological Data Base according to HC-System Approach Yeboah KA, Akpoti K, Kabo-bah AT, et al (2022) Assessing climate change projections in the Volta Basin using the CORDEX-Africa climate simulations and statistical bias-correction. Environmental Challenges 6:100439. https://doi.org/10.1016/j.envc.2021.100439 IPCC 2021 report. https://www.vox.com/22620706/climate-change-ipcc-report-2021-ssp-scenario-future-warming. Accessed 13 Jan 2024 Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-3927808","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":271458989,"identity":"6855041d-c68a-4d4e-8c6f-ea27d24c56b4","order_by":0,"name":"Mukesh Kumar Dey","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA10lEQVRIie3OrwrCUBTH8SMHjmW6ekRxr3BloG2+yhWDRYNFDIaBcKPWPYbJfGSwNPQVtJhNJvHfsFi2gcVwv/FwPvADsNn+MNcNlVzmjzYhUHahItKIpHOKUvQJsSRRon2/ZnAQAhY+f5J4yzVDo1XVTY6wCKDeDPNFZWlmzHtnYhCrCpIhUEvyCULaZTXjNyEGEiDW+YRg3GVNakQZuZcgDox9JUbrjFRMCcKcTE9hKp3XsJ4arIZOIekflpv4NhfPW+/Ox8s1aHtRAflOv5fabDab7feeZLs5JZgKEg0AAAAASUVORK5CYII=","orcid":"","institution":"Indian Institute of Technology Bombay","correspondingAuthor":true,"prefix":"","firstName":"Mukesh","middleName":"Kumar","lastName":"Dey","suffix":""},{"id":271458990,"identity":"ea540e9e-06c5-43db-8c28-c3d375133eb7","order_by":1,"name":"Chandan Kumar Singh","email":"","orcid":"","institution":"National Institute of Technology Raipur","correspondingAuthor":false,"prefix":"","firstName":"Chandan","middleName":"Kumar","lastName":"Singh","suffix":""}],"badges":[],"createdAt":"2024-02-04 13:59:42","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-3927808/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3927808/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":50938613,"identity":"0129fadd-15af-44e6-9d99-af7d5601275f","added_by":"auto","created_at":"2024-02-09 21:41:27","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":120635,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eGroundwater level in the state of Chhattisgarh in the year 2020\u003c/em\u003e\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-3927808/v1/4227092f8eea51b002651b9b.png"},{"id":50938614,"identity":"101295c5-e84c-420d-9d0b-c86e8a07dadf","added_by":"auto","created_at":"2024-02-09 21:41:27","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":152536,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eLithology types of the observation wells\u003c/em\u003e\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-3927808/v1/138e07e6f0f320c113f46ca5.png"},{"id":50938615,"identity":"3f77016f-7dcb-4548-8e31-1c4363bbe36d","added_by":"auto","created_at":"2024-02-09 21:41:27","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":181257,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eLULC change of Chhattisgarh state over the years\u003c/em\u003e\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-3927808/v1/c728d4f70516929fcfa2c372.png"},{"id":50938616,"identity":"def35b02-4355-4c5a-890f-f1d6684708cb","added_by":"auto","created_at":"2024-02-09 21:41:27","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":124172,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eMethodology flowchart of this study\u003c/em\u003e\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-3927808/v1/67938e311bb88e116c00a18d.png"},{"id":50938619,"identity":"372e4e8c-fb6a-45c8-8975-b9c7fc1eeec6","added_by":"auto","created_at":"2024-02-09 21:41:27","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":94803,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eBest fit distribution according to the groundwater level\u003c/em\u003e\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-3927808/v1/ac5319489251ba052ff0caa3.png"},{"id":50938618,"identity":"0e169886-9db6-4226-a35b-bc2610e7d8cc","added_by":"auto","created_at":"2024-02-09 21:41:27","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":139089,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eTime series plot of historical and future annual precipitation of Chhattisgarh state\u003c/em\u003e\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-3927808/v1/70c62f052bb0214c1f77e198.png"},{"id":50938617,"identity":"0e4d263f-aa10-40af-875a-4ac479a3aa68","added_by":"auto","created_at":"2024-02-09 21:41:27","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":136509,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eTime series plot of historical and future annual Max. Temperature of Chhattisgarh state\u003c/em\u003e\u003c/p\u003e","description":"","filename":"7.png","url":"https://assets-eu.researchsquare.com/files/rs-3927808/v1/34fb3eb568d1f02c0b1cd79c.png"},{"id":50938626,"identity":"bedabbaa-a007-418a-a5ac-2c7a3a181d06","added_by":"auto","created_at":"2024-02-09 21:41:27","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":88774,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eTime series plot of historical and future annual Min. Temperature of Chhattisgarh state\u003c/em\u003e\u003c/p\u003e","description":"","filename":"8.png","url":"https://assets-eu.researchsquare.com/files/rs-3927808/v1/70acf695169a55f66a1f3621.png"},{"id":50939053,"identity":"706525b2-e2f7-4390-8af4-7bc824dc7793","added_by":"auto","created_at":"2024-02-09 21:49:27","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":86054,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eActual vs Predicted groundwater level plot for Well ID W200715081320001\u003c/em\u003e\u003c/p\u003e","description":"","filename":"9.png","url":"https://assets-eu.researchsquare.com/files/rs-3927808/v1/7de2b8e706b110e14943eb2d.png"},{"id":50938620,"identity":"aafb830d-355e-46f9-aac7-59a1554b6906","added_by":"auto","created_at":"2024-02-09 21:41:27","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":88275,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eActual vs Predicted groundwater level plot for Well ID W205815080512501\u003c/em\u003e\u003c/p\u003e","description":"","filename":"10.png","url":"https://assets-eu.researchsquare.com/files/rs-3927808/v1/2dd1cd3d2b3954661faeb61e.png"},{"id":50938621,"identity":"fb82086b-717b-4a7a-896c-e331e1400770","added_by":"auto","created_at":"2024-02-09 21:41:27","extension":"png","order_by":11,"title":"Figure 11","display":"","copyAsset":false,"role":"figure","size":104802,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eActual vs Predicted groundwater level plot for Well ID W211245081404002\u003c/em\u003e\u003c/p\u003e","description":"","filename":"11.png","url":"https://assets-eu.researchsquare.com/files/rs-3927808/v1/1ee23b9cdf8db55ab8c7d0f5.png"},{"id":50938624,"identity":"8777230f-ba7c-4079-8a8d-16cc516cb3dc","added_by":"auto","created_at":"2024-02-09 21:41:27","extension":"png","order_by":12,"title":"Figure 12","display":"","copyAsset":false,"role":"figure","size":227363,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003ePredicted groundwater level according to SSP 126 from 2021 to 2100\u003c/em\u003e\u003c/p\u003e","description":"","filename":"12.png","url":"https://assets-eu.researchsquare.com/files/rs-3927808/v1/7862bc6d0873f85005a57e3d.png"},{"id":50938622,"identity":"fad35c4d-4848-410b-b43c-fb333652fd79","added_by":"auto","created_at":"2024-02-09 21:41:27","extension":"png","order_by":13,"title":"Figure 13","display":"","copyAsset":false,"role":"figure","size":215581,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003ePredicted groundwater level according to SSP 245 from 2021 to 2100\u003c/em\u003e\u003c/p\u003e","description":"","filename":"13.png","url":"https://assets-eu.researchsquare.com/files/rs-3927808/v1/690dcf053b3ad34277779f64.png"},{"id":50938629,"identity":"917ca909-c483-4a65-88fe-611443eca106","added_by":"auto","created_at":"2024-02-09 21:41:28","extension":"png","order_by":14,"title":"Figure 14","display":"","copyAsset":false,"role":"figure","size":226245,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003ePredicted groundwater level according to SSP 370 from 2021 to 2100\u003c/em\u003e\u003c/p\u003e","description":"","filename":"14.png","url":"https://assets-eu.researchsquare.com/files/rs-3927808/v1/7029439be0d5906d63c4f30a.png"},{"id":50938630,"identity":"01acdb4a-a892-47c6-ae30-4061af0b38ae","added_by":"auto","created_at":"2024-02-09 21:41:28","extension":"png","order_by":15,"title":"Figure 15","display":"","copyAsset":false,"role":"figure","size":213683,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003ePredicted groundwater level according to SSP 585 from 2021 to 2100\u003c/em\u003e\u003c/p\u003e","description":"","filename":"15.png","url":"https://assets-eu.researchsquare.com/files/rs-3927808/v1/91e0605acd57b3e0f6a681fc.png"},{"id":50938628,"identity":"c1199502-1e6b-4dda-b34d-1b567b8afb78","added_by":"auto","created_at":"2024-02-09 21:41:27","extension":"png","order_by":16,"title":"Figure 16","display":"","copyAsset":false,"role":"figure","size":67917,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eComparison of Groundwater Level of different SSPs with baseline\u003c/em\u003e\u003c/p\u003e","description":"","filename":"16.png","url":"https://assets-eu.researchsquare.com/files/rs-3927808/v1/9218296a6b740823a6d2e07f.png"},{"id":50938631,"identity":"83cea08c-62c0-4c73-9d3d-aaf72a64b6fd","added_by":"auto","created_at":"2024-02-09 21:41:28","extension":"png","order_by":17,"title":"Figure 17","display":"","copyAsset":false,"role":"figure","size":86989,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eComparison of percentage change of Groundwater Level of different SSPs with baseline\u003c/em\u003e\u003c/p\u003e","description":"","filename":"17.png","url":"https://assets-eu.researchsquare.com/files/rs-3927808/v1/d44b04c7b9e26017c6176ab7.png"},{"id":51146976,"identity":"faf6a29b-1138-48fe-967c-dc3d10188886","added_by":"auto","created_at":"2024-02-15 00:54:13","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2568467,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3927808/v1/db637712-5e92-4caa-841e-170aa927ead2.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Forecasting of Groundwater Level Variation Under Changing Climate in Chhattisgarh State Using Deep Learning Technique","fulltext":[{"header":" 1. Introduction","content":"\u003cp\u003eClimate change is the long-term changes of climatic factors. In the 21st century, humans are the most responsible ones behind climate change\u0026nbsp;(Nations 2022). Climate change's impact on natural resources is quite notable. The urgency of the global climate change problem has grown so, investigating climatic conditions and variables has become increasingly popular. The problem of climate change in the world becomes even more critical, so, there has been a growing interest in investigating climate characteristics and factors. Groundwater is one of the most precious natural resources. In India, human activities rely more on groundwater than surface water. Agriculture sectors in India directly depend on groundwater and the ratio of use groundwater and surface water is increasing year by year\u0026nbsp;(NSSO 2014). The vast use of groundwater results in fluctuations in groundwater levels in pre-monsoon and post-monsoon. These abrupt fluctuations in groundwater affect the farmers to estimate their area sowing crops. The precise climate data that can be used to create efficient strategies for both mitigation and adaptation is the greatest method for communities to adjust to changing climatic circumstances\u0026nbsp;(Reddy and Saravanan 2023). Therefore, exact and trustworthy groundwater level (GWL) predictions, offer crucial data on the future availability of groundwater and may be used as the foundation for management choices and plans. Climate models are constantly being updated. Models updated in terms of higher spatial resolution, new physical processes, and biogeochemical cycles\u0026nbsp;(Bader et al. 2008). In 2021, the IPCC released its sixth assessment report, which included CMIP6 climate models.\u0026nbsp;In contrast to its predecessors, the most recent CMIP6 GCMs produce a considerably more accurate picture of Earth's physical processes\u0026nbsp;(Calvin et al. 2023). The CMIP6 models in particular use SSPs to forecast future occurrences\u0026nbsp;(Schlund et al. 2020).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eClimatic factors such as Precipitation and Temperature vary in long-term future different climatic scenarios. Shared Socioeconomic Pathways (SSPs) scenarios are created by an international team of climate scientists and economists to give the climate change research community guidance for doing an integrated, multidisciplinary examination\u0026nbsp;(Riahi et al. 2017). According to the IPCC study, the SSPs have five storylines, including sustainable development, regional competition, inequality, fossil-fueled development, and middle-of-the-road development\u0026nbsp;(UNEP 2022).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eCMIP6 model comes with a certain region-wise bias\u0026nbsp;(Yeboah et al. 2022). We used biased corrected data which is freely available and open for use\u0026nbsp;(Mishra et al. 2020). 13 CMIP6 models are available for India. Each model comes with four SSP scenarios (SSP126, SSP245, SSP370, SSP585) and historical data. This data comes with 0.25\u003csup\u003e0 *\u003c/sup\u003e0.25\u003csup\u003e0\u0026nbsp;\u003c/sup\u003espatial resolution and three parameters i.e., precipitation, t\u003csub\u003emax\u003c/sub\u003e, t\u003csub\u003emin\u003c/sub\u003e. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe artificial neural network (ANN) approach has recently proved its value in forecasting groundwater levels solely based on meteorological variables\u0026nbsp;(Joshi et al. 2020). Recent studies have concentrated on the viability of using machine learning (ML) methods, particularly the branch of ML known as deep learning (DL), in different subfields of hydrology. Some ultramodern engineering uses of ML, such as RL and other innovative methods for applying ML to conventional prediction problems, have not yet been explored. The best approaches to use this geospatial temporal dataset and enhance integrated hydrological system modeling are provided by technologies that belong under the category of geospatial and geo-spatiotemporal AI, such as DL and parallel computing (Ghobadi and Kang 2023). First off, earlier research concentrated on traditional ML models, but the more recently developing DL attention-based models (such as long- and short-term time-series networks (LSTNet), transformer, informer, and conformer) are still in their infancy, especially in the field of WRM\u0026nbsp;(Agarap 2019).\u003c/p\u003e\n\u003cp\u003eCGWB is the main central apex agency in India for groundwater. CGWB takes seasonal readings of groundwater levels all over India. CGWB has various observation wells in India to take reading seasonally. In the generation of Remote Sensing and GIS, CGWB takes groundwater level readings manually. Chhattisgarh states have 1098 observation wells. CGWB dept. Under govt. of India takes groundwater level readings in these wells in 4 seasons in a year- monsoon (June-September), Post-monsoon (October-November), Winter (December-February), and Pre-monsoon (March-May). GW level measurement manually of observation wells at a large scale in Chhattisgarh is very time-consuming and not cost-beneficial at all.\u003c/p\u003e\n\u003cp\u003eIn this study, we investigated both the direct and indirect key drivers for forecasting seasonal groundwater levels. To the author’s knowledge, the inclusion of indirect drivers such as anthropogenic activities and lithology to forecast groundwater levels using machine learning techniques has been scarcely investigated\u0026nbsp;(Obahoundje et al. 2017). To address this issue, the present study explores both the direct and indirect drivers for forecasting seasonal groundwater levels in the Chhattisgarh state, India. We designed a deep learning-based model that considers land cover dynamics, lithological properties, and climatic variables such as temperature and precipitation to forecast seasonal groundwater model.\u003c/p\u003e\n\u003cp\u003eWe use the ANN approach to predict seasonal groundwater levels with high accuracy using precipitation, temperature, land use land cover percentage, and lithology condition as inputs. The calibration and validation of the model were performed using historical climate data. We forecasted and compared the change in groundwater levels for four Shared Socioeconomic Pathways (SSPs) scenarios. This will give a proper idea of how the groundwater level will be in upcoming years and how we can efficiently use it in different sections to maintain the sustainability of groundwater use. This paper presents an innovative method that can be applied to farmers and key stakeholders for effective planning and management of groundwater.\u0026nbsp;\u003c/p\u003e"},{"header":"2.\tData and study area","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\n \u003ch2\u003e2.1 \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eStudy Area Description\u003c/span\u003e: -\u003c/h2\u003e\n \u003cdiv class=\"BlockQuote\"\u003e\n \u003cp\u003eThe area of interest in this study is Chhattisgarh state. Chhattisgarh, state of east-central India. It is bounded by the Indian states of Uttar Pradesh and Jharkhand to the north and northeast, Odisha (Orissa) to the east, Telangana (formerly part of Andhra Pradesh) to the south, and Maharashtra and Madhya Pradesh to the west (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). Its capital is Raipur. Area 52,199 square miles (135,194 square km). Pop. (2011) 25,540,196. The rectangular extent of Chhattisgarh state is 80.24 E, 17.78 N, 84.39 E, 24.10 N. Chhattisgarh state has five major divisions i.e., Surguja division, Bilaspur division, Durg division, Raipur division, Bastar division. Each division has a specific district. According to the latest report, a total of 33 districts are presently available in Chhattisgarh. The central part of Chhattisgarh is our area of interest which mainly consists of 14 major districts. The major districts are- Bastar, Bilaspur, Dhamtari, Durg, Janjgir-Champa, Jashpur, Kanker, Kawardha, Korba, Koriya, Mahasamund, Raigarh, Raipur and Rajnandgaon.\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\n \u003ch2\u003e2.2 \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eData Selection\u003c/span\u003e: -\u003c/h2\u003e\n \u003cdiv class=\"BlockQuote\"\u003e\n \u003cp\u003eThe required data were collected from different websites and the CGWB office which is govt. body of India as shown in Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Tab1\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eData used in this study\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"5\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eSr. No.\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eData\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eComponents\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eResolutions\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eSources\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003e1\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eIMDAA data\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePrecipitation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e(12 km*12 km)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://rds.ncmrwf.gov.in/datasets\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2 m Temperature\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"3\"\u003e\n \u003cp\u003e\u003cstrong\u003e2\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"3\"\u003e\n \u003cp\u003eBias-corrected CMIP6 data\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePrecipitation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"3\"\u003e\n \u003cp\u003e(0.25\u003csup\u003e0\u003c/sup\u003e * 0.25\u003csup\u003e0\u003c/sup\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"3\"\u003e\n \u003cp\u003e\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://zenodo.org/record/3873998#.Y1pQtnZBzIX\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMax. Temperature\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMin. Temperature\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003e3\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eGroundwater level field data\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eWell depth\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eObservation well wise\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eCGWB office, Raipur\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eWater level\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e4\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLULC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLand Use Land Cover Change in Chhattisgarh\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e30 m\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLANDSAT Satellite\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e5\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLithology Condition\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLithology type\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eObservation well wise\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCGWB office, Raipur\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003eGroundwater level (GWL) data from 1995\u0026thinsp;\u0026minus;\u0026thinsp;2020 for 267 observation wells have been used in this study. The maximum depth of the observation wells is 290 m. The well type varies from dug wells and bore wells to tube wells. Central Ground Water Board (CGWB), the main body of India measures groundwater levels all over India. They take groundwater level readings in these wells in 4 seasons in a year- monsoon (June-September), Post-monsoon (October-November), Winter (December-February), and Pre-monsoon (March-May).\u003c/p\u003e\n \u003cp\u003eWe used two climatic variables i.e., Precipitation and Temperature with spatial resolution of 12 Km \u0026times; 12 Km for historical data and 0.25\u003csup\u003e0\u003c/sup\u003e \u0026times; 0.25\u003csup\u003e0\u003c/sup\u003e (\u0026sim;27.5 \u0026times; 27.5 km in the equation) spatial resolution for future climate data. The future climatic data is biased corrected and downscaled, and it is freely available (Mishra et al. \u003cspan class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eThe lithology describes the physical characteristics of the rocks. The central part of Chhattisgarh has 9 types of lithology conditions (Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e). The types are- basalt, charnokite, gnesis, gneissic Complex- basement, granite, limestone, sandstone, schist, and shale. Each lithology has its properties and different hydraulic conductivity values.\u003c/p\u003e\n \u003cp\u003eWe used the hydraulic conductivity value of the lithology types. Lithology is the physical characteristics of rocks and most of them are impermeable, so the water-holding capacity is very low. That\u0026rsquo;s why the hydraulic conductivity value of rocks is very low which is mentioned below (Table \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Tab2\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eHydraulic conductivity value of each Lithology (Widodo et al. \u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e)\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"3\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eSr. No.\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eTypes of Lithology\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eHydraulic conductivity(k)\u003c/p\u003e\n \u003cp\u003em/s\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eI.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGneissic Complex-Basement\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.000002164\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eII.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eShale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.000000411499\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eIII.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLimestone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.00000163791\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eIV.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGneiss\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.000002164\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eV.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSandstone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.00000190922\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eVI.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eBasalt\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0001018519\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eVII.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCharnokite\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.00005956\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eVIII.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGranite\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0000041\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eIX.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSchist\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.000000694444\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003eLand use Land cover outlines the spatial information of distinct categories of the cover over the earth\u0026apos;s surface. It identifies how the area is being used and detects the type of cover. Several types of LULC can be water bodies, bare soil, urban areas, agricultural areas, forest cover, etc. This input data plays a significant role in the amount of flooding in the area (Oxoli et al. \u003cspan class=\"CitationRef\"\u003e2018\u003c/span\u003e). In the current study land use land cover was done by supervised classification. The training data used for supervised classification was the ESRI LULC map for 2020. It was downloaded from the ESRI website. Thirty-year Land Use Land cover was prepared in the interval of 5 years i.e. from 2000 to 2020.\u003c/p\u003e\n \u003cp\u003eFigure \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e shows the land use land cover changes over Chhattisgarh state from year 2000 to 2020. The LULC map is made in a systematic manner using the Colab environment. Raw data for the classification of the area was LANDSAT Satellite imageries (Manandhar et al. \u003cspan class=\"CitationRef\"\u003e2009\u003c/span\u003e). Making use LANDSAT 9 LULC for the year 2020 was prepared, while LANDSAT 8 was used for the year 2015, and for the years 2010, 2005, 2000, and 1995 LANDSAT 5 imageries were used. The Land use land cover was prepared in the Colab environment using GEE (Google Earth engine). Colab is a cloud-based Jupyter notebook that runs in the personal Google workbook. It can access the web and the internet. Python code can be written and executed in such an environment. A geospatial processing service is called Google Earth Engine. With the help of Earth Engine and the Google Cloud Platform, you can carry out geospatial processing at a scale. Earth Engine is designed to create a platform that is interactive for large-scale geospatial algorithm development, making significant advancements in addressing global concerns using big geospatial datasets. Making use of these two tools land use land cover was prepared and the area was classified into seven classes namely water, cropland, bare land, forest area, built-up area, shrubs, and other vegetation.\u003c/p\u003e\n\u003c/div\u003e"},{"header":"3.\tMaterials and Methodology","content":"\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\n \u003ch2\u003e3.1 \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eMethodology\u003c/span\u003e: -\u003c/h2\u003e\n \u003cp\u003eFigure \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e outlines the modified approach to forecasting seasonal groundwater levels using a deep learning-based technique. Contrary to the conventional approach that makes use of limited input, we introduced a modified approach that combines both climate data and indirect drivers. The rainfall and temperature are the main climatic factors that drive the model. The output is based on climatic variables. The input variables are rainfall and temperature as a time series. Based on the input variable the model is calibrated and then after the validation of calibration, it predicts the output variable.\u003c/p\u003e\n \u003cp\u003eThe modified approach, which is the main method of this study, is based on a combination of climatic factors, lithology conditions land use, and land cover changes over the years. We took five variables as our input variables. The input variables are- rainfall, temperature, hydraulic conductivity of the lithology condition, and the land use land cover. The input variables are converted into usable format from their raw format using Python in the pre-processed step. The input variables are then sorted as a time-series pattern to supply the model. Then we separate the input variable into two parts- features and levels. The feature is the individual, independent variables serving as. To create predictions, prediction models require characteristics. A feature engineering approach may also be used to create new features from existing ones. The label is the finished product. The output classes refer to the labels. Data scientists refer to a collection of samples that have been assigned to one or more labels when they talk about labeled data. The time series is now Split into two parts- the training set and the test set. A piece of our real dataset called the \u0026quot;training set\u0026quot; is fed into the deep learning model to help it find and understand patterns. This develops the model. Usually, training data is bigger than testing data. This is because we want to provide the model with as much information as we can for it to identify and learn useful patterns. When our dataset\u0026rsquo;s data are supplied to a machine learning algorithm, this recognizes patterns in the data and concludes. The test set analyses the effectiveness and development of your training algorithms and alters or optimizes them for better outcomes. Two primary standards apply when testing data. It should be a true representation of the dataset and be big enough to produce accurate predictions. Test data offer a last-minute, practical verification of an unknown dataset to show that the deep learning algorithm was successfully trained.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\n \u003ch2\u003e3.2 \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eModel development and performance\u003c/span\u003e: -\u003c/h2\u003e\n \u003cdiv class=\"BlockQuote\"\u003e\n \u003cp\u003eThe proposed approach is applied as a filter to increase the efficacy of hydrological models. Almost all scientific and technical fields have effectively used deep learning as a tool (Rozos et al. \u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e). Machine learning models have advanced in hydrology from basic feed-forward networks used for short-term forecasting to complicated models that can consider even the static properties of catchments, mimicking the hydrological experience (Ghobadi and Kang \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e). According to recent studies, machine learning models are more reliable and effective than conventional hydrological models (both conceptual and physically based), often exceeding them (Ghobadi and Kang \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e). An indirect approach to take advantage of ML in hydrological applications is to use it as a tool for pre-processing the data or post-processing the results of the standard hydrological models (Panahi et al. \u003cspan class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eTo build these types of models we first need to understand the nature of data. In our dataset, there are 2 types of data are there- features and labels. The label data is the most important data to calibrate and validate our results. To calibrate the model, it\u0026rsquo;s particularly important to set the hyperparameter accurately. The hyper-parameters were adapted based on trial-and-error methods. We used various hyper-parameters to ensure that the model works accurately. First, we choose the activation in this model as an \u0026lsquo;Adam\u0026rsquo; optimizer with a learning rate of 0.01. The batch size and epoch size are 33 and 2000, respectively. We chose the activation function as a Rectified linear unit (relu) which depends upon the best fit curve of the groundwater level in each well.\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv class=\"BlockQuote\"\u003e\n \u003cp\u003eFigure \u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003e Shows the best fit curve is lognormal. Similarly, we check it for all the observation wells. The largest wells show the results in between lognormal and gamma. According to this distribution function, we choose our activation function as \u0026lsquo;relu\u0026rsquo;. Although some wells show normal distribution.\u003c/p\u003e\n \u003cp\u003eThe total GWL data set is taken for 1995\u0026ndash;2020 i.e., 26 years (26 \u0026times; 4\u0026thinsp;=\u0026thinsp;104 seasons). This dataset is split into two parts: the training set is taken 15 years i.e., from 1995 to 2009 and the testing data is set for the remaining years from 2010 to 2020, 11 years. Lastly, we forecasted the GWL up to the end of the 21st century (2100).\u003c/p\u003e\n \u003cp\u003eThe calibrated and validated results are checked by Nash-Sutcliffe Efficiency (NSE) and squared Pearson\u0026rsquo;s correlation coefficient (R\u003csup\u003e2\u003c/sup\u003e) matrix.\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv id=\"Equa\" class=\"Equation\"\u003e\n \u003cdiv class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\u003cimg src=\"data:image/png;base64,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\"\u003e\u003c/div\u003e\n \u003c/div\u003e\n \u003cdiv class=\"BlockQuote\"\u003e\n \u003cp\u003eAfter the usable NSE and R\u003csup\u003e2\u003c/sup\u003e value we used it for future forecasting. Good NSE and R\u003csup\u003e2\u003c/sup\u003e value means the actual value of groundwater and the forecasted value is nearly the same for these wells.\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e"},{"header":"4.\tResults and Discussions","content":"\u003cp\u003eWe applied our model to the future data. We sorted our future data according to the observation well and different SSP conditions. We all know that groundwater is a highly fluctuating earth property. As the climate changes very rapidly, the trend of groundwater level of also changing. Figure \u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003e shows that the annual trend of precipitation will change drastically in the future among all the SSPs. The blue line in the Fig. \u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003e is the historical precipitation in the Chhattisgarh state. The red dotted line is the worst SSP condition, and we also observed it in the below fig. The maximum change in precipitation is shown in this SSP condition. The Figs. \u003cspan class=\"InternalRef\"\u003e7\u003c/span\u003e and \u003cspan class=\"InternalRef\"\u003e8\u003c/span\u003e shows the annual tread of max. and min. is over the Chhattisgarh state. The max. temperature is above 37\u003csup\u003e0\u003c/sup\u003eC at the end of the 21st century. The min. temperature recorded at 27\u003csup\u003e0\u003c/sup\u003eC at the end of the 21st century. All the SSPs in fig. show increasing order.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCalibration and validation performance\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe take 60 percent of our input data as training data and the remaining 40 percent of our data as test data. We also computed the actual and model forecasted (2010\u0026thinsp;\u0026minus;\u0026thinsp;2020) GWL time series with future forecasting up to the end of the 21st century (2021\u0026thinsp;\u0026minus;\u0026thinsp;2100) for all the wells.\u003c/p\u003e\n\u003cp\u003eFigure \u003cspan class=\"InternalRef\"\u003e9\u003c/span\u003e, Fig. \u003cspan class=\"InternalRef\"\u003e10\u003c/span\u003e, \u003cstrong\u003eand\u003c/strong\u003e Fig. \u003cspan class=\"InternalRef\"\u003e11\u003c/span\u003e summarizes the validation performance of our model. The below figures not only match the shapes but the peak points also capture very well. These wells show an accuracy level above 0.80 which is a good one.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSSP wise forecasting\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOur goal is to forecast the future groundwater level in different SSP scenarios. In our study, we take 4 SSP scenarios i.e., SSP 126, SSP 245, SSP 370, and the worst condition SSP 585. The different SSP conditions have different carbon emissions, depending upon it has different forecasted precipitation and temperature values. We apply our calibrated model in these SSP scenarios. We want to visualize the change in groundwater level in different future scenarios. This will give a proper idea of how the groundwater level will be in upcoming years and how we can efficiently use it in different sections to maintain the sustainability of groundwater use.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSSP 126\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe SSP 126 is one of the good future scenarios. According to the SSP126 scenario, a sustainable future society will develop with significant economic growth, investments in health and education, and strong governance (Siabi et al. \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e). This scenario anticipates a 1.8 C increase in global average temperature by 2100, which is in line with the Paris Agreement\u0026apos;s goals of limiting warming to below 2.0 C (Rogelj et al. \u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e).\u003c/p\u003e\n\u003cp\u003eFigure \u003cspan class=\"InternalRef\"\u003e12\u003c/span\u003e shows the forecasted groundwater level of Chhattisgarh state of two different wells according to SSP 126. Fig shows the different season-wise groundwater levels in different colors. The green line which is season 3 (June- August) which is monsoon season where rainfall is maximum shows groundwater level minimum among all the seasons. Season 1 (December - January) and season 4 (September- November) groundwater levels in the range of 5 to 8 m. Sometimes the middle of the 21st century and season 4 touched a groundwater level of 10 m which is quite exceptional. The maximum level of groundwater forecasted by the model is in season 2 (February - May) which is the summer season where the precipitation is minimal. The forecasted groundwater level is in the range of 9 m to 11 m. The mean value of this season is around 10 m. The orange line of the above plot does not vary abruptly apart from 2060 to 2068. Here the groundwater level is around 9 m. The SSP 126 is the optimistic evolution of future society (O\u0026rsquo;Neill et al. \u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e). So, the groundwater level is also in a certain range which is in the controllable situation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSSP 245\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe SSP 245 scenario is called the \u0026ldquo;\u003cstrong\u003emiddle of the road\u003c/strong\u003e\u0026rdquo; scenario (Fricko et al. \u003cspan class=\"CitationRef\"\u003e2017\u003c/span\u003e). This scenario is consistent with the climate initiatives taken by different nations to uphold their current obligations to cut emissions. The total radiative forcing is anticipated to increase in SSP 245 to around 4.5 Wm \u003csup\u003e\u0026minus;\u0026thinsp;2\u003c/sup\u003e before declining after 2070 (Siabi et al. \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e). This leads to increased heat extremes and catastrophic flooding from precipitation extremes also predicted to cause a drop in food production throughout the planet.\u003c/p\u003e\n\u003cp\u003eFigure \u003cspan class=\"InternalRef\"\u003e13\u003c/span\u003e shows the forecasted groundwater level according to SSP 245. Here the groundwater level changing treads are somehow different from SSP 126 (IPCC 2000; harrisson 2018). The green color line which is season 3 shows groundwater level in the range of 2 to 6 m. This is the minimum level among all the other SSP scenarios. In this season the groundwater level fluctuates most. This is the monsoon season where the precipitation is also not uniform, this can be one of the reasons for this fluctuation. The summer seasons show the level in the range of 9 m to 10 m. In this season, in the last couple of years of the 21st century, the level slightly increases. Those years show the groundwater level above 10 m. The mean groundwater level of this season in SSP 245 is around 10 m. As this SSP is in the middle of all the SSP scenarios, the groundwater level also not touching extremely high in this SSP scenario.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSSP 370\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe SSP 370 scenario is the 3rd most common scenario where the burning of coal we are increasing slowly. Countries have a propensity to withdraw from international collaboration and concentrate on their economic aims in the SSP 370 scenario. Methane, aerosols, and other gases that trap heat other than carbon dioxide are anticipated to reach high concentrations (Meinshausen et al. \u003cspan class=\"CitationRef\"\u003e2011a\u003c/span\u003e). By 2100. According to this scenario, there will be 12.6\u0026nbsp;billion people on the planet. Radiative forcing levels are predicted to increase to roughly 7.0 Wm\u003csup\u003e\u0026minus;\u0026thinsp;2\u003c/sup\u003e at the end of 2100. Sea levels will rise drastically in this SSP. According to this SSP, the sea level will be between 46 and 74 cm (Riahi et al. \u003cspan class=\"CitationRef\"\u003e2017\u003c/span\u003e).\u003c/p\u003e\n\u003cp\u003eFigure \u003cspan class=\"InternalRef\"\u003e14\u003c/span\u003e shows the forecasted groundwater level according to SSP 370. The groundwater levels in both season 1 and season 4 overlapped with each other like the previous two SSP forecasting. The range of fluctuating groundwater water levels is between 6 m to 8 m. In the year between 2066 to 2069, the groundwater level suddenly decreased to 4 m for both observations. From Fig. \u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003e we can observe that in those years the precipitation drastically increased. Figures \u003cspan class=\"InternalRef\"\u003e7\u003c/span\u003e and \u003cspan class=\"InternalRef\"\u003e8\u003c/span\u003e show that the trends of maximum and minimum temperature are also in decreasing order. That is why the groundwater level also decreased. The groundwater level in the summer season is in the range of 8 m to 10 m. But at the end of 2100 years, the groundwater level shows slowly increasing order. From the year 2086 to 2100 the groundwater level shows an increase in order. At the end of the 21st century, it touches almost 11 m.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSSP 585\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn the SSP 585 scenario, humanity generally does nothing to stop climate change but continues to worsen it. In this scenario, burning coal, natural gas, and oil will all contribute to global economic growth. Radiative forcing levels are predicted to increase to roughly 8.5 Wm \u003csup\u003e\u0026minus;\u0026thinsp;2\u003c/sup\u003e in this scenario (Meinshausen et al. \u003cspan class=\"CitationRef\"\u003e2011b\u003c/span\u003e). Large-scale coastal flooding and incredibly damaging storms are expected in this scenario. Parts of the world will become uninhabitable as a result, particularly during the warmest times of the year. This will be the worst scenario. The average temperature rise will be 3.7\u003csup\u003e0\u003c/sup\u003e C in this SSP. According to the IPCC report, 2\u003csup\u003e0\u003c/sup\u003e C temperature rise is recognized as the threshold at which climate change becomes dangerous. The sea level rise will be 0.63 m. These effects will lead to the adaptation cost. In this SSP scenario, the adaptation cost will be much higher than other SSP scenarios.\u003c/p\u003e\n\u003cp\u003eFigure \u003cspan class=\"InternalRef\"\u003e15\u003c/span\u003e shows the forecasted groundwater level according to SSP 585. In this SSP scenario, all the seasons show increasing trends. The groundwater level in monsoon seasons is also quite high and increasing trends. In the monsoon season, the level is in the range of 2 m to 8 m. The groundwater level fluctuates a lot in this season. In season 1 (December- January) the groundwater level touched around 10 m which is the maximum in all the SSPs. In this SSP the climatic variable also shows an increase throughout the period, that\u0026rsquo;s the reason the groundwater level also increased. The summer season shows the highest groundwater level. The average groundwater level will be 11 m in this SSP scenario. In season 2 the groundwater level touched almost 12 m at the end of 2100. This will lead to groundwater scarcity in the upcoming years.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBaseline comparison\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe season-wise comparison for each SSP with a baseline line is shown in Fig. \u003cspan class=\"InternalRef\"\u003e16\u003c/span\u003e. At The baseline we take the last observed groundwater level. The last observed groundwater level was taken in the year 2020 by CGWB. We compare the groundwater level from the line with each SSP scenario and we can visualize the changes. The historical groundwater maximum in season 2 (February -May). Apart from SSP 585, all the other SSPs i.e., 126, 245, and 370 show the forecasted groundwater level is maximum in Season 2 itself. In SSP 585 season 1 shows the maximum ground water level. The orange bar in the plot denotes the season 3 groundwater level. It is the minimum in baseline and all the SSPs scenarios. The peaks are much higher in SSP 585 compared to all other SSPs. This will be more climate change-affected scenarios among all the others. In this SSP the precipitation pattern also drastically changes with max. and min. temperature. This affects the groundwater level that we can observe from the above plot.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePercentage change with baseline\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFigure \u003cspan class=\"InternalRef\"\u003e17\u003c/span\u003e shows the percentage change of groundwater level in different SSP scenarios season-wise. These changes we compare with the Baseline mean groundwater level in the year 2020. We can see the changes in the Fig. \u003cspan class=\"InternalRef\"\u003e17\u003c/span\u003e. The groundwater levels are drastically changing over the year. The maximum changes are noted in SSP 585 and SSP 245. In these two scenarios, the groundwater level changes by more than 350%. In SSP 585 all the seasons show groundwater increasing trends of more than a minimum 150%. The season4 shows an increasing trend in all the SSPs. In this season the precipitation and temperature show increasing in order and the pattern of precipitation also changes. These changes a visualized in Fig. \u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003e. These changes cause more Vulnerability in the upcoming future years.\u003c/p\u003e"},{"header":" Conclusions","content":"\u003cp\u003eThe prime objective of this study was to introduce a modified approach using a machine learning technique that considers the climate variables along with additional key factors. \u0026nbsp;This study has focused on assessing the impacts of climate-changing scenario and their effects on groundwater levels. In this study, we forecasted groundwater levels using ANN models.\u0026nbsp;This deep learning-based model forecasts accurately in those observation wells where the historical data is available and minimum data gaps are there. The data gaps also play a significant role in predicting the groundwater level.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThis study proves that the aquifer condition land use and land cover change also play and key role in the fluctuation of the groundwater level other than the rainfall and temperature. The more permeable aquifer condition shows satisfactory results other than rocks that are impermeable or semi-permeable. The changes in land use and land cover are also going to be responsible for groundwater level decline in the future. The increase of built areas and decreasing order of vegetation also affect the groundwater level. This affects the recharge area which can also handle groundwater level changes.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eDifferent SSP conditions show different groundwater level values. Although it is confirmed that the groundwater level for the upcoming future scenario will decline year by year. Each year the season 3 period means the monsoon period (June to August) shows groundwater in the range of 3 to 6 m. The summer season (February to May) shows groundwater in the range of 8 to 12 m. The other two seasons - season 1 (December to January) and season 4 (September to November) show groundwater levels in the range of 6 to 9 m. One of the main important observations of our study is that- the more permeable lithology condition of the observation wells shows better results than impermeable or semi-permeable aquifer-type observation wells. This study may in the future extend to climate change impacts on extreme events such as floods and droughts in terms of their frequency and occurrence should be analyzed. In this study we only used one model which was selected by literature review and earlier works have been done in central India, based on that we selected this model.\u003c/p\u003e\n\u003cp\u003eAs we forecast groundwater levels seasonally. This forecasting will help the farmers to irrigate their land according to the available groundwater in these regions. This forecast will be very beneficial for farmers, and key stakeholders such as the state water resources department, Central Groundwater Board (CGWB), etc. to maintain the recharge-discharge ratio by planning annual groundwater discharge for sustainable development. The farmers may get an appropriate information idea for irrigating their land based on the seasonal availability of groundwater for irrigation purposes. This enhanced information about groundwater level forecasts may help in short-term and long-term planning and management of the groundwater.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAuthor Contributions:\u003c/strong\u003e Both authors contributed to the study at all levels and original draft preparation. \u003cstrong\u003eMukesh Kumar Dey\u003c/strong\u003e: Methodology, Visualization, Writing – original draft; \u003cstrong\u003eChandan Kumar Singh\u003c/strong\u003e: Conceptualization, Writing – review, editing final draft.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u003c/strong\u003e No funding was used in this research.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of Data and Materials:\u003c/strong\u003e Data will be available upon reasonable request.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDeclarations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests:\u0026nbsp;\u003c/strong\u003eThe authors declare no competing interests.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eAgarap AF (2019) Deep Learning using Rectified Linear Units (ReLU)\u003c/li\u003e\n \u003cli\u003eBader D, Covey C, Gutowski W, et al (2008) Climate Models: An Assessment of Strengths and Limitations. Climate Models: An Assessment of Strengths and Limitations\u003c/li\u003e\n \u003cli\u003eCalvin K, Dasgupta D, Krinner G, et al (2023) IPCC, 2023: Climate Change 2023: Synthesis Report. Contribution of Working Groups I, II and III to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change [Core Writing Team, H. Lee and J. Romero (eds.)]. IPCC, Geneva, Switzerland. Intergovernmental Panel on Climate Change (IPCC)\u003c/li\u003e\n \u003cli\u003eFricko O, Havlik P, Rogelj J, et al (2017) The marker quantification of the Shared Socioeconomic Pathway 2: A middle-of-the-road scenario for the 21st century. Global Environmental Change 42:251\u0026ndash;267. https://doi.org/10.1016/j.gloenvcha.2016.06.004\u003c/li\u003e\n \u003cli\u003eGhobadi F, Kang D (2023) Application of Machine Learning in Water Resources Management: A Systematic Literature Review. Water 15:620. https://doi.org/10.3390/w15040620\u003c/li\u003e\n \u003cli\u003eharrisson \u0026nbsp; thomas (2018) Explainer: How \u0026lsquo;Shared Socioeconomic Pathways\u0026rsquo; explore future climate change. In: Carbon Brief. https://www.carbonbrief.org/explainer-how-shared-socioeconomic-pathways-explore-future-climate-change/. Accessed 13 Jan 2024\u003c/li\u003e\n \u003cli\u003eIPCC (ed) (2000) Emissions scenarios: summary for policymakers;a special report of IPCC Working Group III$Intergovernmental Panel on Climate Change. Intergovernmental Panel on Climate Change\u003c/li\u003e\n \u003cli\u003eJoshi N, Rahaman M, Thakur B, et al (2020) Assessing the Effects of Climate Variability on Groundwater in Northern India. World Environmental and Water Resources Congress\u003c/li\u003e\n \u003cli\u003eManandhar R, Odeh IOA, Ancev T (2009) Improving the Accuracy of Land Use and Land Cover Classification of Landsat Data Using Post-Classification Enhancement. Remote Sensing 1:330\u0026ndash;344. https://doi.org/10.3390/rs1030330\u003c/li\u003e\n \u003cli\u003eMeinshausen M, Smith SJ, Calvin K, et al (2011a) The RCP greenhouse gas concentrations and their extensions from 1765 to 2300. Climatic Change 109:213\u0026ndash;241. https://doi.org/10.1007/s10584-011-0156-z\u003c/li\u003e\n \u003cli\u003eMeinshausen M, Smith SJ, Calvin K, et al (2011b) The RCP greenhouse gas concentrations and their extensions from 1765 to 2300. Climatic Change 109:213. https://doi.org/10.1007/s10584-011-0156-z\u003c/li\u003e\n \u003cli\u003eMishra V, Bhatia U, Tiwari AD (2020) Bias-corrected climate projections for South Asia from Coupled Model Intercomparison Project-6. Sci Data 7:338. https://doi.org/10.1038/s41597-020-00681-1\u003c/li\u003e\n \u003cli\u003eNations U (2022) What Is Climate Change? In: United Nations. https://www.un.org/en/climatechange/what-is-climate-change. Accessed 10 Mar 2023\u003c/li\u003e\n \u003cli\u003eNSSO (2014) Statistics on Indian Economy and Society. https://www.indianstatistics.org/irrigation.html. Accessed 4 Jan 2023\u003c/li\u003e\n \u003cli\u003eObahoundje S, Ofosu EA, Akpoti K, Kabo-bah AT (2017) Land Use and Land Cover Changes under Climate Uncertainty: Modelling the Impacts on Hydropower Production in Western Africa. Hydrology 4:2. https://doi.org/10.3390/hydrology4010002\u003c/li\u003e\n \u003cli\u003eO\u0026rsquo;Neill BC, Tebaldi C, Van Vuuren DP, et al (2016) The Scenario Model Intercomparison Project (ScenarioMIP) for CMIP6. Geosci Model Dev 9:3461\u0026ndash;3482. https://doi.org/10.5194/gmd-9-3461-2016\u003c/li\u003e\n \u003cli\u003eOxoli D, Ronchetti G, Minghini M, et al (2018) Measuring Urban Land Cover Influence on Air Temperature through Multiple Geo-Data\u0026mdash;The Case of Milan, Italy. ISPRS International Journal of Geo-Information 7:421. https://doi.org/10.3390/ijgi7110421\u003c/li\u003e\n \u003cli\u003ePanahi M, Sadhasivam N, Pourghasemi HR, et al (2020) Spatial prediction of groundwater potential mapping based on convolutional neural network (CNN) and support vector regression (SVR). Journal of Hydrology 588:125033. https://doi.org/10.1016/j.jhydrol.2020.125033\u003c/li\u003e\n \u003cli\u003eReddy NM, Saravanan S (2023) Extreme precipitation indices over India using CMIP6: a special emphasis on the SSP585 scenario. Environ Sci Pollut Res. https://doi.org/10.1007/s11356-023-25649-7\u003c/li\u003e\n \u003cli\u003eRiahi K, van Vuuren DP, Kriegler E, et al (2017) The Shared Socioeconomic Pathways and their energy, land use, and greenhouse gas emissions implications: An overview. Global Environmental Change 42:153\u0026ndash;168. https://doi.org/10.1016/j.gloenvcha.2016.05.009\u003c/li\u003e\n \u003cli\u003eRogelj J, Den Elzen M, H\u0026ouml;hne N, et al (2016) Paris Agreement climate proposals need a boost to keep warming well below 2 \u0026deg;C. Nature 534:631\u0026ndash;639. https://doi.org/10.1038/nature18307\u003c/li\u003e\n \u003cli\u003eRozos E, Dimitriadis P, Bellos V (2022) Machine Learning in Assessing the Performance of Hydrological Models. Hydrology 9:5. https://doi.org/10.3390/hydrology9010005\u003c/li\u003e\n \u003cli\u003eSchlund M, Lauer A, Gentine P, et al (2020) Emergent constraints on equilibrium climate sensitivity in CMIP5: do they hold for CMIP6? Earth Syst Dynam 11:1233\u0026ndash;1258. https://doi.org/10.5194/esd-11-1233-2020\u003c/li\u003e\n \u003cli\u003eSiabi EK, Awafo EA, Kabo-bah AT, et al (2023) Assessment of Shared Socioeconomic Pathway (SSP) climate scenarios and its impacts on the Greater Accra region. Urban Climate 49:101432. https://doi.org/10.1016/j.uclim.2023.101432\u003c/li\u003e\n \u003cli\u003eUNEP W (2022) IPCC report: Climate Change 2022\u003c/li\u003e\n \u003cli\u003eWidodo L, Cahyadi T, Notosiswoyo S, Widijanto E (2016) Application of Clustering System to Analyze Geological, Geotechnical and Hydrogeological Data Base according to HC-System Approach\u003c/li\u003e\n \u003cli\u003eYeboah KA, Akpoti K, Kabo-bah AT, et al (2022) Assessing climate change projections in the Volta Basin using the CORDEX-Africa climate simulations and statistical bias-correction. Environmental Challenges 6:100439. https://doi.org/10.1016/j.envc.2021.100439\u003c/li\u003e\n \u003cli\u003eIPCC 2021 report. https://www.vox.com/22620706/climate-change-ipcc-report-2021-ssp-scenario-future-warming. Accessed 13 Jan 2024\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":true,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Climate change, Groundwater, Forecasting, Deep learning, Neural network, CMIP6, Land Use Land Cover","lastPublishedDoi":"10.21203/rs.3.rs-3927808/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3927808/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eGroundwater is the most precious natural resource in modern days. India is the largest consumer of groundwater globally, with over 25% of the world's groundwater extraction. Climate change affects the groundwater level both in direct and indirect ways. Recently developed machine learning approaches have led to the consideration of selected climate variables that can govern the groundwater dynamic. The inclusion of indirect key drivers such as anthropogenic activities and lithology to forecast groundwater levels using machine learning techniques is not well understood.\u0026nbsp; This paper aims to consider both the direct and indirect key drivers for forecasting seasonal groundwater levels. In this context, a modified approach based on a deep learning model has been formulated that considers land cover dynamics, lithological properties, and climatic variables such as temperature and precipitation. The model was calibrated and validated to forecast seasonal groundwater levels for four Shared Socioeconomic Pathways (SSPs) scenarios.\u0026nbsp; The results show that the accuracy level, R\u003csup\u003e2\u003c/sup\u003e is 0.86 which is acceptable. Overall, the results obtained broadly correspond to an acceptable degree of accuracy. \u0026nbsp;The proposed methodology is applicable for seasonal groundwater level forecasting and can be useful to farmers and key stakeholders.\u003c/p\u003e","manuscriptTitle":"Forecasting of Groundwater Level Variation Under Changing Climate in Chhattisgarh State Using Deep Learning Technique","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-02-09 21:41:22","doi":"10.21203/rs.3.rs-3927808/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"afd7b396-23ec-4df7-a577-359c3e55485a","owner":[],"postedDate":"February 9th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-02-15T00:46:05+00:00","versionOfRecord":[],"versionCreatedAt":"2024-02-09 21:41:22","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-3927808","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3927808","identity":"rs-3927808","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}