Drainmod Model Adaptation for Developing Recommendations Water Management in the Tertiary Block of Tidal Lowland Agriculture

The primary key to successful agricultural cultivation is maintaining the groundwater level to fulfill crop water requirements, particularly during the dry season. Field study was conducted in Tidal reclamation area of section 25 at Sugihan Kanan, Bandar Jaya Village, Air Sugihan, Ogan Komering Ilir District of South Sumatra. The DRAINMOD computer model was used to simulate water levels in dry and wet climatic conditions. The principal measured parameters are soil hydraulic conductivity and drain spacing, as well as daily rainfall data. The simulation results showed that the research area belongs to the rainfed type, and the main objective of water management is to retain water and determine some efforts to increase the groundwater level through pump irrigation in the dry season. The application of pump irrigation was applied to the plant entering the generative phase. The pump irrigation was provided to distribute water into the quarter and worm (micro) channels. The effect of this application caused the groundwater level to approach about 30 cm below the soil surface, while groundwater level in areas without pump irrigation facility was in the range of 50-60 cm. Besides efforts to increase the water table, liming is still required in order to increase production. Lime application of 1 Mg ha had a significant effect on increasing production. Corn production with this treatment could produce 5 Mg ha, while non-treated land areas only produce 2-3 Mg ha.


INTRODUCTION
Extensive land clearing of tidal lowland areas for agricultural enterprise in South Sumatra was started in 1969. About 400,000 ha of tidal lowland had been reclaimed through the transmigration program. Banyuasin area has shown successful effort in rice production at South Sumatra with a magnitude higher than 40 percent of total rice production (Desperate 2017). However, this successful effort had not followed by other agricultural areas because the planting index at these areas was less than 200%. One relatively less productive area is Sugihan Kanan, with a rice production level of fewer than 3 Mg.ha -1 , and most of its agricultural land had one planting index (Imanudin et al. 2016). The soil is classified as acid sulfate soil, with high acidity and Alumunium content (Fahmi et al. 2014). Aluminum toxicity has produced an alteration of biochemical and physiological reactions of plants and then to their crop productivity. Decreasing root growth is one of the initial and most evident symptoms of Al-toxicity. The nutrient availability was also decreasing due to bond reaction Al-P (Koesrini et al. 2014). On the other hand, Increasing levels of aluminum and iron in the soil solution also cause a decrease in the quality of water in the canals, soil pH and water pH dropping to a range of 2.3-3. Then the water channeled should be fluch out and replaced by freshwater from the rain or tidewater (Sukitprapanon et al. 2019).
The key success for land management at tidal lowland areas is how to manage soil water level at elevation level required by crop's root zone (Imanudin et al. 2010;Bakri et al. 2015). According to Fahmi et al. (2014), water status availability in the root zone is the main factor determining the thriving agriculture in acid sulfate soil. Rainwater was an essential water resource, mainly in acid sulfate soil. The pyrite oxidation process during a dry period made groundwater a highly toxic element August 4, 2020;Revised June 18, 2021; Accepted 20 June 2021 and very acidic. Freshwater from the tide is not possible for land irrigation. Thus, to fulfill water crop requirement mainly achieved by rainwater (Imanudin et al. 2019). Therefore water retention in channels is vital to the block technique (Nurzakiah et al. 2017). Water in the canal is managed at a 60-90 cm depth to create groundwater at 10-30 cm (Herawati et al. 2020). With the controlled drainage system water in the canal is still available to keep groundwater table in the root zone and always above the pyrite layer. So that pyrite oxidation can be avoided in the soil profil (Sasirat et al. 2019).
Another critical factor is the availability of structures and infrastructure of water management networks to facilitate the land leaching process and water flooding in channels. Inadequate drainage facility frequently results in toxic elements accumulation and high soil acidity (Bakri et al. 2015). In addition, not all tidal lowland areas can provide good water quality from a secondary channel for supply (Hartoyo et al. 2010;Megawaty et al. 2012). Many channels had carried acid water or saline water during the dry season, which should be prevented from entering tertiary channels (Hairmansis et al. 2017;Tafarini and Yazid 2018). Watergates at the tertiary channel must be installed to control the drainage and water retention process (Lasmana et al. 2017;Imanudin et al. 2019).
The construction of field operation planning requires daily water table data during rain and dry seasons, either at dry or wet climate conditions (Imanudin et al. 2010;Negm et al. 2016). Daily water table data so far is frequently unavailable, difficult to measure, and costly. Therefore, the DRAINMOD computer model can aid in presenting the dynamics condition of the daily water table (Masoud et al. 2021;Wahba et al. 2018;Ashkan et al. 2020). This data can be used to construct weekly water management operational plans in the field for crop cultivation (Imanudin et al. 2011). According to Skaggs et al. (2012), statistical analysis shows acceptable results where the simulation model with daily water level prediction results calibrated with field data shows Nash-Sutcliffe (EF) modeling efficiency values are 0.68 and 0.72, the daily drainage rate (EF = 0.73) and 0.49), and monthly drainage volume (EF = 0.87 and 0.77). The simulated result was a high correlation between predicted and measured (Wan et al. 2009;Malakshahi et al. 2020). Proper drainage planning by the DRAIMOD simulation model also depended on the quality of data input. Physical and Hydrological parameters such as rainfall and soil hydraulic conductivity were essential data input. (Negm et al. 2014;Ashkan et al. 2020). The model was also successful for developing land drainage design (Sojka et al. 2019). The model can provide input on the proper use and operation of drainage network infrastructure (Ashkan et al. 2020). The model could provide better predictions of groundwater table depths under shallower drainage systems. Moreover, produce management tools to minimize environmental issues in the agriculture field (Davoodi et al. 2019). The DRAINMOD model is excellent drainage modeling for estimating the depth of the groundwater table (Davoodi et al. 20019). The estimated groundwater level from simulation and modeling results shows the value of r 2 = 0.93 (Ashkan et al. 2020). An adaptation model for the tropical area has been carried out in tidal lowland reclamation areas in Banyuasi, South Sumatra. The simulation result found r 2 0.83. Hydrological parameters are essential factors that will have a better statistical analysis (Imanudin et al. 2011). This paper will present the results of water table dynamics obtained from DRAINMOD modeling and operational planning implemented during planting season. Field adaptation was conducted by corn planting at dry season. The recommended DRAINMOD model for water supply in tertiary channels will be implemented by rainfall harvesting method (passive approach) and active approach through water pumping from the tertiary channel during the dried condition. The objectives of this field study are to construct the operational model for a water table control for agriculture purposes.

MATERIALS AND METHODS
This research was implemented at the tidal lowland area of tertiary plot-4 at Bandar Jaya Village, Jalur 25, Air Sugihan Subdistrict, Ogan Komering Ilir District ( Figure 1). It was conducted from May to October 2017. The research site is classified as C typhology land in which it can not receive water through high tidal irrigation because high tidal water can not enter into this land. Water from secondary channel only capable of filling tertiary channel.
Materials used in this study are soil samples, corn seeds, fertilizers, pesticides, crops protecting plastics, and chemicals for soil analysis in the laboratory. Equipment used in this study are consisted of a piezometer, wells (perforated PVC pipe), 12 inch PVC pipe, elbow, marking board, water pass, meter, soil bor, excavation tube (bailer), stopwatch), flap watergate made of fiber, 10 inch PVC pipe, digital camera, and agricultural equipment. For evaluation of water status at tertiary block and land, drainage planning was conducted through computer simulation using DRAINMOD 5.1 software (Skaggs et al. 2012).
The conceptual model to develop DRAINMOD is based on the water balance analysis within vertical soil column unit per surface area. The calculation starts from the impermeable layer up to the soil surface between drainage channels (Skaggs 1978). Calculation of water balance within soil profile for a time period of dt is mathematical can be expressed as follows : Where Va  xi is the change of soil air volume (cm), F is infiltration (cm), ET is evapotranspiration (cm), D is lateral flow (negative sign indicates drainage flow and a positive sign indicates subsurface irrigation conditions) in cm, Ds is side seepage flow (positive sign indicates capillary flow upward) in cm, P is precipitation (rainfall) in cm, RO is surface flow (cm), and S  xj is the change of surface water storage. As an illustration, the drainage profile system in DRAINMOD can be seen in Figure 2. Rainfall inputs in the DRAINMOD model are hourly rainfall, and maximum and minimum  temperatures, which are read from weather data and water balance conducted every hour. Input parameters for DRAINMOD are soil physical characteristics consisting of impermeable layer depth, soil hydraulic conductivity value, and climate data consisting of temperature and daily rainfall. Water network information has consisted of the depth and distance amongst channels. A simulation was conducted to determine soil water table dynamics at dry and wet climate conditions.
Observation of water table monitoring was extended to a period prior to dry season in May 2017. The corn seeds used in this study were Pioneer P21 varieties. To support the land drainage system, farmers developed a microchannel up to 30 cm with a distance between channels of 8 to 10 m.
There are two treatments were used in this study: firstly, crop without lime addition and water source only from watergate retention; supply water only from the water capillarity movement; Secondly, a pump added crops with 1.5 Mg.ha -1 lime and water supply from irrigation. In order to provide water in the tertiary channel for pumping operation, farmers had conducted temporary water retention (dam) by using canvas plastics. The pumping operation was conducted at the crop's generative phase. Irrigation used in this study was furrow irrigation in which water is delivered to fill collector channel, and water from this channel subsequently was distributed to fill microchannel. Irrigation was applied once a week.
The computer model DRAINMOD was used for predicting the daily water table as the effect of deference climatic conditions. Data input of the DRAINMOD simulation model were hydraulic conductivity, soil water retention, land drainage system (drain spacing, depth, and width of channel), and crop physiological data. The scenario for the model operation was constructed based on different climatic conditions. The analysis method for soil water status under the root zone was calculated using the SEW-30 concept. The calculation of SEW-30 is based on the Sieben formula (Skaggs et al. 2012). This concept is used to determine the condition of soil water excess (cm day -1 ) during the crop growth period. The concept of water excess 30 cm above the root zone has an objective to evaluate fluctuation height of soil water table during a wet period (rainy season) within tidal lowland agriculture area (Imanudin et al. 2018). The value of water excess 30 cm above the root zone can predict soil water excess during crop growth. The equation used for this calculation is as follows: where Xi is the soil water table at (days too), with i is the first day and n is the number of days during crop growth. The DRAINMOD model calculates the hourly value of SEW-30 cm instead of the daily value so that the calculation of SEW-30 value is more accurate and can be formulated by using the following equation: where Xj is the soil water table at the end of respective hours, and m is the total hours during the crop growth. Water table position with a critical limit of 30 cm is done with consideration value of 30 cm below the soil surface is selected because most of the food crops will experience physiological disturbance if the soil water table is drop down from 30 cm point or increase upward from 30 cm point from the soil surface. If soil water is far from the critical value of 30 cm or closes to the soil surface, it will create excess water conditions (Negm et al. 2014). This condition is applied for non-rice food crops. On the contrary, rice crop is withstanding water flooding condition and will experience stress if soil water is below 30 cm depth zone or even below 20 cm depth zone (Imanudin et al. 2011;Imanudin et al. 2019).

Soil Physical Characteristics
Soil physical characteristics in the study area are shown in Table 1. Soil has high porosity and low bulk density in the top layer. Meaning that soil had relatively high total pore space resulting in relatively high water holding capacity. The decrease of soil bulk density is followed by the decrease of total pore space at deeper soil layers. The water movement means that on the rotting zone area, the water movement was easy to drain and submerge by gravity forces.
The soil layer above 60 cm showed the increase of clay content so that soil bulk density can be classified into mineral soil. The main limiting factor is soil acidity level than can be classified as acid soil with a 4-5. It is also reported by Koesrini et al. (2014) when pH <5, the Aluminum toxicity is an increase in soil. It is the main stress factor for plant growth in tidal lowland agriculture.
Soil textural class is shown by the ratio of clay, sand, and loam fractions ( Table 2). Although soil fraction composition had occurred, soil texture at the study area is qualitative can be classified into silt loam class at 0-60 cm and clay texture in the layer of 60 cm below the soil surface. Clay fraction tends to increase with the increase of soil depth.
Clay fraction with the magnitude of 7.6 % at soil layer depth 0 -30 cm had increased to 21.6 % at a depth of 60 cm. The increase of clay fraction below soil surface had produced a beneficial effect in which land can store water at depth above 60cm so that soil water depth at this layer is not quickly depleting. In addition, soil water contribution at this layer through capillary movement is significant to supply crop water requirements. According to Zipper et al. (2015), the contribution of groundwater through capillary rise is strongly dependent on soil texture, and increasing clay fraction on the soil more provides water by capillary rise than soil having sandy soil. Added by Gao et al. (2017), at groundwater depth 1 m below the soil surface, the capillary upward was supplied about 41% of the crop evapotranspiration, which means that the water supply was required to fulfill crop water requirements when the water table started to drop at 70 cm.

Computer Simulation of DRAINMOD in estimating soil water table
DRAINMOD is a drainage model capable of estimating the depth of soil water table for swamp lowland and peatland areas. The main inputs for this model are consist of rainfall, crop evapotranspiration, soil hydraulic conductivity, and drainage network characteristics data. The water flow condition within the soil is assumed to be constant (steady condition) in the Drainmod model (Skaggs et al. 2012).
In the case of reclaimed tidal lowland area at Sugihan Kanan, some model inputs have consisted of soil hydraulic conductivity of 0.34 m.day -1 , impermeable layer depth of 2 m, the distance between tertiary channels at a current condition in the field was 250 m, initial soil water depth is assumed 10 cm below the soil surface and average channel depth of 1.5 m. The constructed scenario can be seen in Table 3. The simulation was conducted at two climatic conditions: typical climate (average rainfall) and dry rainfall due to Elnino. The simulation stages have consisted of simulation on the existing network condition without control effort followed by simulation on watergate operation. Computer simulation results of DRAINMOD for wet and dry climate conditions are shown in Figure 3. Simulation results showed that the soil water table variation at the condition of the initial to the final rainy season had a safe value in land firing probability. However, simulation results also showed that over drain had been occurred on land, and they were indicated by a quick drawdown of soil water table elevation in case of no rainfall occurrence for more than one week period. The soil water table fluctuation results either at normal or dry rainfall conditions (Elnino) showed a similar trend in terms of soil water table conditions for the January-May period.
The land was still in safe condition from fire hazards during this period. Although the soil water table drops into critical depth value (80 -100 cm), rainfall still occurred so that soil water elevation could increase again close to 30 cm depth. The vulnerable condition started to occur at days of 140 (entering of May) in which rainfall was decreased, and soil water elevation was continuously drop exceeding the critical limit (-80cm). In this period, pyrite oxidation could also haven and create a decreasing pH in the root zone.
On the other hand, the soil has a pyrite layer around 60-70 cm below the soil surface. When the groundwater table level drops to 100 cm, an oxidation layer will occur. This process would produce high soil acidity and increase the solubility of iron and aluminum content, which may be harmful to cropsreported by Koesrini et al. (2014) that applying lime with 2.0 Mg ha -1 is still required for increasing pH in the soil. Added by Fahmi et al. (2014), the lime application was efficient in combination with good land preparation through intensive soil leaching and intermittent waterlogging. It was also reported by Ar-Riza et al. (2015) that the main problem is challenging to get fresh water for the leaching and flushing process. Those for managing acid sulfate soil are still required lime application to increase pH. According to Aksani et al. (2018), increased productivity of tidal lowland rice cultivation was achieved by 10 Mg ha -1 rice straw compost, and NPK fertilizers that should be applied are 315 kg urea ha -1 , 135 kg SP-36 ha -1 , and 90 kg KCl ha -1 , Mean that water availability is not only a single factor for increase land productivity.
Rainfall was average in 2014; soil water elevation continuously dropped to the lowest point of -115 cm in September and started to increase when entering October, although it was still within  1  14  27  40  53  66  79  92  105  118  131  144  157  170  183  196  209  222  235  248  261  274  287  300  313  326  339  352  365 Days Water tablet depth (cm) the critical depth limit. The safe condition again occurred in November, whereas the dry climate condition (Elnino) in 2015 showed that soil water elevation dropped to a depth of -118 cm, which occurred in September. However, soil water elevation was still at a critical depth value up to December due to minimal rainfall. Since soil had high porosity, farmers were reluctant to open tertiary channels because rainfall water was assumed to be insufficient to supply the soil water table and channels. Therefore, water pumping from the tertiary channel should be made to deliver water into quarterly channels, which requires stepping detention during high tidal water in the tertiary channel.
Simulation results from Figure 1 showed an open drainage system with the distance between channels of 250 m and channel dimensions as follows: upper width of 2.0 m and lower width 1.2 m, as well as the depth of 1.2 m, was excessive in terms of dimension size. This channel dimension is a high capacity to drain water. Therefore, water losses were rapid, indicated by a water drawdown of up to 40 cm in case of no rainfall for two weeks.
Simulation using a shallow drainage system was subsequently tried to determine the proper channel dimension at peatland to prevent excessive water losses. The depth of the tertiary channel was only 1 m with a bottom width of 30 cm. Simulation result (Figure 2) showed that soil water table drawdown was not as fast as the existing channel dimension.
Simulation results of shallow drainage showed that soil water table condition in 2015 (Elnino) was within a safe zone up to June. The control measure can be done in the June period. Because the drainage system had already been built and it was impossible to conduct filling activity, the choice was to conduct water retention immediately using a controlled dam system. Therefore, a computer simulation of DRAINMOD was implemented at the initial stage. The simulation was also conducted to determine the impact of water retention within the tertiary channel (Figure 4). Water retention simulation would be done during the least rainfall occurrence period in May. Simulation results showed that soil water table elevation could be increased up to a depth of 30-40 cm below the soil surface during May-June. Soil water table elevation was located at 60 cm below the soil surface during the dry season of July-November. Land at this condition was relatively safe from fire hazards. This condition would be achieved if soil water table elevation in the tertiary channel was not dropped more than 40 cm from the soil surface.
The condition of this controlled drainage system was found in the field during October, where the soil water table was located at a safe zone, i.e., 40-50 cm below the soil surface. This condition also prevents pyrite oxidation, although pumping irrigation supply was needed to fulfill crop water requirements. Pump facilities are highly needed if farmers cultivate corn at B and C land typology areas during the dry 1  14  27  40  53  66  79  92  105  118  131  144  157  170  183  196  209  222  235  248  261  274  287  300  313  326  339  352  : Water retention option, : No option (Uncontroll).
season. The pump was only operated for one month during the seed filling phase (generative period), and irrigation was applied once a week. A water retention structure was highly needed because the land area at the study location had high water losses. Control structures are needed to support network operation as water retention (rainfall harvesting). The dam pattern equipped with an overflow structure is the proper model for the tertiary channel ( Figure 5). Farmers currently use sandbags to withstand water losses. The effort to elevate soil water table elevation can be made by water pumping from the tertiary channel into the quarterly channel, which was subsequently delivered further into the collector and micro (worm) channels.
Control drainage is the best option for water management in the area study. The farmer keeps the water in the tertiary canal minimum at 50 cm above the bottom level of the canal. By maintaining the water level in the tertiary 50-60 cm than the groundwater table purposed above the pyritic layer during the dry season.

Field Adaptation of Water Dynamics at Dry Season Condition (July 2017)
Water table elevation was hourly observed to determine soil water table dynamics, which is affected by the fluctuation of water surface in the channel. Results of this hourly observation can be seen in Figure 6. The average value of water table depth was 37 to 41 cm below the soil surface, and this magnitude was ideal for the growth of the corn crop. Hourly observation of the soil water table showed a decrease of water surface followed by a Figure 5. Hydraulic structure in tertiary canal for water retention and control drainage in tidal lowland type C.  : Water table in secondary canal,  : water table in tertiary,  : ground water table. decrease in the secondary channel in the tertiary channel. In contrast, the increase of water surface in the secondary channel was followed by the increase in the tertiary channel. The change of soil water table did not follow this condition. The soil water table movement was relatively stable from a maximum depth condition of -50 cm into a minimum depth condition of -33 cm. The movement of the soil water table was highly affected by rainfall. The insignificant difference in water table depth at the tertiary block was due to water availability in the tertiary channel. Therefore, the horizontal movement of water (see page) was practically slight and can be neglected as long as water along the tertiary channel was available at the height of 50-60 cm.
The relationship between water level elevation in the tertiary channel and soil water table as a result of water retention impact (Tabel 4) showed that water level elevation in a tertiary channel on May-July was located at 70 cm depth and water level elevation in the secondary channel was located at 168 cm depth. This condition can produce soil water table elevation at 30 cm below the soil surface.

The network operation and corn growth (PS2)
The network operational guidelines in the tertiary channel can be summarized in Table 5. The operation was conducted each month according to crop growth phases.
Crop evapotranspiration requirement at dry season entering September could not be fulfilled if solely rely on capillary water movement from soil water table depth of 50 cm. Capillary water could only supply 45% of crop evapotranspiration requirements at this condition, so another water supply was needed. Farmers used a water pump to fill the soil water table (Figure 7). A significant increase of soil water table was observed on corn cultivation which was supplied through pump irrigation. Water can be elevated with a magnitude of 20 cm. The soil water table on land without pump irrigation was located 40-50 cm below the soil surface, whereas the soil water table on land supplied with pump irrigation was located at about 30 cm (Figure 8). If the soil water table was located at a depth of 30 cm below the soil surface, then crop evapotranspiration requirement can be fully supplied from capillary water movement. Yang et al. (2011) reported that the capillary rise could not support the crop water requirement when the groundwater table was below 2.5m. Moreover, when the groundwater table maintenance is at 50 cm, irrigation is not required.
Corn production achieved using maximum technological inputs consisting of improvement of irrigation system, and soil quality showed good yield with a magnitude of about 3 Mg.ha -1 . However, this production level was still lower than corn production   (Imanudin and Bakri 2014). The yield of corn would be an optimum underwater table at 50-60 cm below the soil surface (Bakri and Bernas 2015). Since the water table dropped by more than 100 cm, the performance of corn growth decreased as reported by Liu and Luo (2011) in Australia that the significant contribution of the groundwater table ranges between 40-150 cm. It was contributed more than 65% of the potential evapotranspiration. However, capillary water rise fulfilled the crop's water requirement under 40-50 cm.

CONCLUSIONS
Hydrotopographical characteristics at the study area were classified into C class in which land area does not receive the overflow from high tidal water. However, water availability in the channel is capable of controlling the soil water table.
Komputer model DRAINMOD computer model can be used to estimate fluctuation of the soil water table at several climatic conditions. Simulation results can be utilized to construct a monthly water management plan.
The main objective for water management at the tertiary level was water retention due to physical soil conditions characterized by high porosity and high hydraulic conductivity. Therefore, the operational model was to retain water during the rainy season (rainfall harvesting) and control drainage during the second crop (corn).
Water retention in the tertiary channel at a depth of 70 cm could maintain soil water table elevation at 30 cm below the soil surface. Soil water table elevation should not be a drop below 50 cm in order to prevent pyrite oxidation.