Prediction of groundwater trends for irrigation in Northern Bangladesh

Groundwater trends affect the domestic, agricultural, and industrial prospects of a region. The study area is Bogura, a northern region of Bangladesh, located on the Pleistocene terrace of the Bengal Basin. The aquifer consists of medium-to-coarse sand, located at a depth of 4.66 – 42.68 m; groundwater is scarce during dry seasons. The water table (WT) time-series data for 2007 – 2019 were used for forecasting and characterizing present and future groundwater conditions using existing numerical simulations. The annual groundwater budget for discharge and storage was 2,772 and 2,442 Mm 3 , respectively. Thus, the annual scarcity of groundwater was 330.4 Mm 3 (13.5%), excluding the surface water contribution of 10 Mm 3 (0.4%). The present spacing of deep tube wells (DTWs) and shallow tube wells (STWs) was 744 and 372 m, respectively. Currently, the DTW spacing ranged 744 – 800 m; however, the STW spacing of 250 – 372 m is higher than the set distance. Hence, further installations of STWs were strictly disallowed for irrigation. WT declined by 1.0 m in the last 13 years, i.e., 0.07 m or 1.2% decline rate per annum, causing water scarcity in the region during the peak period in the dry season (June – February), thus affecting irrigation and limiting agricultural production.

tables (WTs) of subsurface water are decreasing in shallow aquifers in Asian mega deltas such as the Ganges-Brahmaputra-Meghna (GBM) basin owing to the large seasonal variation in the monsoon rainfall in Bangladesh.
The penetration rates of the high Barind soil were 0.0015 md À1 for wet land and 0.0075 md À1 for dry land (MacDonald & Partners ). The WT is as much as 15 m below the ground surface, and groundwater fluctuation differs from 1.25 to 16.15 m (Asaduzzaman ). In 1990, the total amount of groundwater withdrawn was 8,806 Mm 3 against a reserve of 21,088 Mm 3 for several agricultural inputs in Bangladesh (MPO ). The irrigation coverage was 1.52 × 10 6 ha during 1982-1983, and it increased to 3.79 × 10 6 ha in 1996-1997. Groundwater also contributes 70% of the water used for irrigation and 90% of the drinking water in the country (Matin et al. ). The current available geological subsurface information indicates that most of the aquifers occur between depths of 30 to 130 m, and the country has heterogeneous aquifer systems (Zahid & Ahmed ). Further, alluvial aquifers are the proven sources of groundwater in Bangladesh, and the groundwater is annually refilled via rainfall and flooding and through the systematic recycling of water nationwide, except in Dhaka Megacity, where there is an imbalance between the asymmetrical recharge-discharge. Approximately 80% of the agricultural land is irrigated using groundwater after the start of the crop season in Bangladesh, and groundwater scarcity arose 42% each year during the operation of STWs in the dry season in northwest Bangladesh (Shahid & Hazarika ). In Barind Tract of Bangladesh, prospective groundwater recharge decreased by 85%, whereas the rest was moderate. In this region, only 8.6% of the total average annual rainfall (1,685 mm) infiltrates into the subsurface and ultimately contributes to the recharge of groundwater (Adham et al. ).
Net recharge was increased approximately 0.005-0.015 m/ year in many region of Bangladesh during 1985-2007 for increased groundwater withdrawal for irrigation as well as consumptive use, but net recharge marginally decreased from À0.0005 to À0.001 m/year, where groundwater irrigation was low; that is, <30% of total irrigation at the same period (Shamsudduha et al. ). Since the commencement of groundwater irrigation in the mid-1980s, a positive impact was observed in cropping patterns, cropping intensity, and national intensity, crop selection, and cultivation yields practices. Besides, the WT has decreased continuously at rates of 0.4 and 0.22 m/year in the wet and dry seasons, respectively (Rahman & Mahbub ). Generally, the rain season starts in May and ends in September, and there is low or no rainfall during the rest of the year; thus, the maximum rainfall is observed in June-August. The maximum WT is observed in July-September because of rainwater infiltration into the subsurface, and the minimum WT is observed in March-May at the time of irrigation. The WT declines continuously owing to overdrawing of groundwater for irrigation in Chapai-Nababganj in northwest Bangladesh (Hasan et al. ). The annual groundwater recharge and discharge were estimated to be 106.41-244 Mm 3 and 93.77-291 Mm 3 , respectively, in Naogaon region in northwest Bangladesh; these data reveal that there existed a balance between the annual recharge and withdrawal up to 1993; however, since 1993, the discharge exceeded the recharge. In all, 23.99-42.08 Mm 3 of groundwater was discharged using STWs and DTWs, and the rest of the groundwater is discharged by natural seepage (Reza et al. ). The groundwater depletion rate ranges from 0.00 to 0.03 m/year based on the mean depth of phreatic surfaces; further, the maximum depth of the phreatic surface varies from 0.012 to 0.15 m/year (Abdullah ). The depth of WT and rainfall penetration are slowly declining in most wells, and the present trends of decline for the WT in Bogura district suggest that the depth of WT may double by 2060 in some regions (Hasanuzzaman et al. ). Barind, the drought prone area in Bangladesh, has a typical dry climate and high temperature except for the winter season, beginning from mid-June to October, whereas the rainfall and temperature ranges are 1,500-2,000 mm and 4 C-44 C, respectively, in the northern region of Bangladesh (www. bmda.gov.bd).
However, groundwater hydrology is a very complex process and is not limited to merely understanding the groundwater situation, zoning maps, temperature, humidity, rainfall, weather, climate, and aquifer properties because of the difficulties in characterizing an area based on only a few studies. Therefore, groundwater hydrology is a continuous process of providing solutions when studying a problem for a specific region. The aquifer in the Bogura region is in a poor conditionduring the dry season, water is insufficient to meet the demands of both irrigation and human consumption; hence, efforts are being made to improve the availability of groundwater resources. Although research on groundwater in northern Bogura of Bangladesh is scanty, most focus on groundwater assessments (Abdullah ; The main objective of the study is of interest to the hydrological community, particularly the Water Users Association (WUA) involved with groundwater resources, especially for irrigation. First, the aquifer is mapped using lithological data of boreholes. Second, the balance between groundwater recharge-discharge and storage is estimated. Third, the safe well spacing for installing the boreholes is determined.
The result can provide insight into the need for digging or abstaining from digging new wells, selection of the device for extracting groundwater, and operability of different modes of the water-lifting equipment. Furthermore, the results will be a handy tool for decision-makers and planners working not only in the study area but also on national and global scales.

METHODS
In this study, the present and future groundwater tendencies were predicted using some recognized empirical equations.
The focus was to obtain a clear understanding of groundwater resources, including the balance between rechargedischarge and storage. The methodological steps are described in the following flowchart: This section provides a broad stepwise description based on appropriate simulation to rationalize the projected scenario of groundwater resources.

Study location
The target area of this study was Bogura in northern Bangladesh ( Figure 1). Geographically, the study area lies between longitudes 89 , 21 min, and 0 s east and from 24 and latitudes 46 min to 48 north. It is called the gateway to north Bangladesh. It is approximately 200 km from the north of Dhaka, the capital of the nation.

Geology of the study area
Seven types of geological group formations are exposed across Bangladesh (Figure 1). Of these, two types of geological formations were found in the study area. These formations include the Pleistocene terrace (Madhupur Clay) and alluvium, flood plain deposits, delta plain deposits, and stream deposits. The mappable body of rock of the Pleistocene terrace is part of the Dihing formation. This enormous formation has aquifer characteristics. The formation contains clays of various colors, such as mediumgrain molted clay and yellow and gray clays, and occasional pebbly sandstone. The rock formations are partly associated in most of the locations, and similarly, alluviums are loosened materials deposited on the flood plains, delta basins, river beds, lakes, or estuaries. Alluviums comprise sand, silt, gravel, cobbles, and boulders along with organic materials that are converted into fertile soil. Generally, the term 'alluvium' is constrained to size-sorted fine sediments (sand, silt, and clay) and unlithified riverine deposits. In addition, coarse-grained and very coarse-grained sand (brown and gray) is found most of the alluvium in the study area. The color of sand is transformed from yellowish and brownish to gray.

Sample size
Lithologies for 145 boreholes were analyzed for mapping aquifers and to prepare a groundwater lithological inventory to identify the soil formation types. Precisely, 11 stations were selected for monitoring the wells ( Figure 2) (one from each station or subunit), and the time-series primary data of WT were collected for the last 13 years (2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018)(2019). The location of the selected monitoring wells is listed in Table 1.

Rockworks software
Rockworks allow computer modeling based on the end-user specifications. The basic strategies of the software are the creation of a borehole database using the analytical results for various physical and chemical properties as a function of the depth. The database also generates images, such as block diagrams, cross sections, and fence diagrams, to check the legitimacy and geological reasonability of the modeling. It uses two-dimensional (2-D) and three-dimensional (3-D) images with three feature levelsbasic, standard, and advanced. It is often difficult to determine the capabilities, operational features, and limitations of a particular groundwater modeling code from the documentation, or even without accurately running the code for situations relevant to the region for which a code is to be selected owing to the incompleteness, poor organization, or incorrectness of code documentation. Thus, this software was used in the study to map the aquifer using lithological data.

Estimation of annual groundwater storage
Groundwater refers to the water present in the pore spaces of soil. That part of the rock or unconsolidated deposit formation is called an aquifer when it can harness an effective quantity of water. Groundwater storage is the product of the area of the aquifer, depth or thickness of the WT fluctuation, and specific yield (Raghunath ): where ΔS is the groundwater storage, A is the area of aquifer in the study area (m 2 ), ØH is the depth or thickness of the WT fluctuation (m), and S y is the specific yield (%). The depth of the WT fluctuation is the difference between its maximum and minimum levels in the WT throughout the year in the given region. In this study, groundwater fluctuations were measured at 11 monitoring wells. They were not constant at every location, and hence, a vertical aquifer was expected. Further, the specific yield (S y ) is the ratio of the thickness of an equivalent layer of water of infinite spatial extent that is required to produce the measured change in gravity (in other words, it is the change in groundwater volume per unit area) by the change in WT elevation.
The Equation can easily assess groundwater discharge by comparing storage, shortage, or steadiness of groundwater.

Groundwater recharge
Groundwater recharge is a hydrological process where water passes downward from the surface deeper into the ground. It is the amount of water that enters the saturated zone; that is, the enduring WT in an aquifer. The main source of recharge is rainfall, which may enter the soil directly to reach the groundwater zone. Aquifer recharge refers to the amount of water that may be available in the long term for extraction. Groundwater recharge, also referred to as deep drainage or deep percolation, is calculated for areas that receive abundant rainfall. The data of where R is the groundwater recharge (mm) and P is the annual rainfall (mm).
(c) Datta's formula (Datta et al. ): where R is the groundwater recharge (cm) and P is the annual rainfall (cm).

Groundwater discharge and surface water contribution
Groundwater discharge is the term used to describe the movement of groundwater from the subsurface to the surface. Again, surface water contribution is the sense used to the water coming from the surface (canals or rivers).
Annual discharge and surface water contribution can be easily determined to ascertain the number of wells (DTWs, STWs, and LLPs), average pumping capacity, and time period of operation. The annual discharge and surface water contribution of groundwater for irrigation in the study area is estimated using the following equation (Asaduzzaman ): where V a is the annual groundwater discharge (Mm 3 ), N is the number of tube wells operated in the area, Q is the pumping capacity of a well (m 3 h À1 ), and t is the operation period (h/year).

Safe well spacing
This is generally defined as the maximum area of the resource basin that can be drained efficiently and economically by a well. The prevailing spacing between wells was determined for the study area by considering a uniform areal distribution of wells and entire utilization of the safe yield of the groundwater basin on the basis of the available groundwater recharge. Safe yield is defined as the amount of water that can be withdrawn from the groundwater basin without producing an undesired effect. The annual recharge must not exceed the total groundwater extracted. The well spacing clearly indicates the distribution of water-lifting devices installed in an irrigation field. The relation of well spacing to recharge, discharge, and pumping period is described by the following equation (Chowdhury & Wardlaw ): where S is the well spacing (m), Q is the well discharge in (L/S), t is the pumping period in (days/year), R g is the groundwater recharge in (mm/year), and N is the number of operating well (h/day).

Growth rate and WT trends
For computing the WT growth rates, the following exponential trend line formula is used.
where Y is the dependent variable (WT; m); t is the independent variable (month-year); a is the intercept, and b is the absolute growth rate. Thus, b is the growth with a ratio scale multiplied by 100 and expressed as a percentage of decrease. To verify the significance of the estimated regression, the following equation for the F-test, where null hypothesis H o : b ¼ 0 was used (Gujarati ): where R 2 is the explained sum of squares and (1 À R 2 ) is the unexplained sum of squares.
Direct field supervision Some data were collected through direct field supervision with consent from end users, beneficiaries, and managers handling groundwater resources in the study area.

RESULTS AND DISCUSSION
The study results are clearly systematically stated and described in this section considering the applied methods. In addition, the maximum, minimum, and average aquifer  Table 2.
From the data in Table 2,  fluctuation, ØH, is 3.37 m (from Table 2
The average groundwater recharge in Table 3    supplementary irrigation is performed for the seasons of Aus and Aman cultivation; that is, July-August and December-January, respectively, if needed).
From Equation (5), the number of wells, average pumping capacity, time period of operation, annual discharge, and surface water contribution are obtained as listed in Table 4.
The annual groundwater storage and annual ground-  Table 4, a total of 273 LLPs are used for contributing surface water with an amount of 9.90 Mm 3 (0.41%).

Safe well spacing
To calculate the well spacing, the following parameters were

Period toward groundwater fluctuations
The list in Table 5

WT trends
The annual exponential value was computed using the trend line for the 11 stations of the northern region of Bogura, Bangladesh, as listed in Table 6. The x-axis denotes the number of years, and the y-axis represents year; the WT is reflected in the exponential value y and regression R 2 . The trend lines shown in Figure 8 were computed using   Equation (7) or Equations (8) and (9), which are presented in Table 6.
The data in Table 6 show that the WT fell by 1. year. This result can be considered in future groundwater user plans.
The study results can be significant for decision-makers, planners, and researchers involved in groundwater management at nationwide and worldwide scales.