INDIAN INSTITUTE OF TECHNOLOGY ROORKEE Deep Percolation from Surface Irrigated Water Intensive Crop Fields By K.S., Hari Prasad (CED, IIT Roorkee, India) Hatiye, Samuel D. (WRIE, Ethiopia) C.S.P. Ojha (CED, IIT Roorkee, India)
Presentation Outline Introduction The problem/concern Objective of the study Literature Methods and materials Experimental Works Mathematical modelling Results and Discussions Conclusion Scope for future work 2
Introduction Fresh water is mainly consumed for the purposes of agricultural, domestic and industrial water needs. Agriculture is by far the largest consumer of fresh water of the globe; that is, water put to irrigate a cropland to produce crops. Mainly in developing countries, More than 80% of fresh water withdrawals goes for agricultural water input (FAO, 2004). 3
Introduction... Due to limitation in resources, surface methods of irrigation are usually practiced in developing countries. In particular, flooding way of water application is implemented in water intensive crops such as paddy(rice) and berseem fodder crops. Large areas are in paddy and berseem cultivation in many parts of the world, on the other hand. 4
Introduction... Rice is a staple food grain for nearly half of the world population. Paddy field is a major/largest consumer of water in the irrigated agriculture. 5
Introduction... There is also an increasing demand for berseem fodder production due to increasing demand for dairy products. Berseem also needs frequent irrigation due to its shallow root depth. But the resource base (water) is limited. Less supply????? More demand. 6
Introduction In agricultural water use, processes such as Deep percolation, seepage, evaporation and runoff are taken as unproductive water losses. Deep percolation refers to the water that flows beyond the crop root zone of a given crop (Wang et al., 2012; Bethune et al., 2008; Ma et al., 2013; Huang et al, 2003).
Problem Statement Deep percolation phenomena from frequently irrigated fields such as paddy and berseem fields seriously diminishes irrigation efficiency, jeopardise proper water management and minimize water productivity. Further it can cause environmental havoc by carrying agricultural based residues and chemicals (surface and/or groundwater pollution). Groundwater level rise and hence water logging and secondary salinization.
Problem Statement. Specifically, deep percolation from water intensive crops in relatively permeable soils needs more attention. Most available studies deal with deep percolation under puddled root zone conditions and ignoring the un-puddled field situations where most farmers practice irrigating their paddy and berseem.
Problem Statement... So far ( The Gap) Very few studies were conducted on deep percolation from paddy and berseem fields covering different regimes of water application and employing drainage type lysimeters. Only little understanding about deep percolation under unpuddled field conditions and different seasons exits. 10
Objective of the Study The main objective of the present study is to estimate deep percolation from surface irrigated water intensive crops such as paddy and berseem fodder fields using the water balance and physically based models while employing drainage type lysimeters. 11
Materials and Methods Experimental Program Laboratory Experiments (Soil physical and hydraulic) Field Experiments(Soil, crop, irrigation monitoring) Deep Percolation Estimation Models Water balance Model Physically based Model 12
Simple Water Balance Model Spatially lumped (root zone) and temporally distributed has been used (Allen et al., 1998; Ochoa et al. 2007; Abrahaoa et. al, 2011). ( P + I + SP i n + GW ) ( RO + DP ET S where P is precipitation, I is Irrigation, SP in and SP ot are seepage/lateral inflow and outflow respectively from the root zone, GW is the capillary rise from groundwater, RO is surface runoff, DP is deep percolation, ET is evapotranspiration and S is change in soil water storage. + + SP ot ) = 13
Physically Based Model The one dimensional Richards (1931) Equation (Liu et al., 2014; Tan et al., 2014) as used in HYDRUS- 1D package is (Simunek et al. 1998): = KK(ΨΨ) + (ΨΨ) SS(zz, tt) where θ is the moisture content, ΨΨ is the pressure head, z is the vertical coordinate usually taken positive upwards, t is the time coordinate, K is the hydraulic conductivity of the soil and S(z,t) is the sink term representing root water uptake. 14
ΘΘ = 1 1+ αα v ΨΨ nn v Constitutive Relationships θθ ΨΨ Relationship:- mm for ΨΨ 0 1 for ΨΨ > 0 ΘΘ = θθ θθ r θθ s θθ r where αα v and nn v are unsaturated soil parameters with m = 1 (1/nn v ) for nn v > 1 ; and ΘΘ is the effective saturation defined as where θθ s = Saturated moisture content; and θθ r = Residual moisture content of the soil. K -θθ Relationship:- mm 2 KK θθ = KK satt ΘΘ 1 2 1 1 ΘΘ 1 mm for ΘΘ < 1 = KK sat for ΘΘ = 1 where KK sat is saturated hydraulic conductivity 15
Experimental Program Laboratory Experiments Laboratory experiments consisting of determination of soil, crop and soil hydraulic parameters were conducted. These include: Soil bulk and particle density, soil texture and soil hydraulic characteristics. These are presented in the following slides. 16
Experimental Program... Laboratory Experiments ρ b = M V t s n =1 ρ ρ b d ρ d = M V s s 17
Experimental Program... Laboratory Experiments 18
Experimental Program... Laboratory Experiments Sample 1(0-30cm) sample 2 (30-60cm) sample 3(60-80cm) sample 4(80-100 cm) sample 5(100-140cm) Average Grain Size Distirbution Curve Grain Size Boundary According To ASTM colloids clay Silt Sand 0.001 0.005 0.075 4.75 0 0.0001 0.0010 0.0100 0.1000 1.0000 10.0000 Grain size (mm) 100 90 80 70 60 50 40 30 20 10 19 Percent Finer, %
Experimental Program... Laboratory Experiments Sample No Depth (Below GL),cm Percent Sand (%) Percent silt (%) Percent Clay (%) Soil Class (USDA) 1 0-30 69.10 26.95 3.83 Sandy Loam 2 30-60 69.20 25.52 5.30 Sandy Loam 3 60-80 68.10 26.18 5.77 Sandy Loam 4 80-100 73.60 22.15 4.23 Sandy Loam 5 100-140 65.40 26.87 7.71 Sandy Loam Average 0-140cm 69.10 25.53 5.37 Sandy Loam 20
Experimental Program... Laboratory Experiments Depth below ground level (cm) Bulk density (g/cm 3 ) Particle density (g/cm 3 ) Sand (%) Silt (%) Clay (%) Soil Class (USDA) Porosity 0-30 1.58 2.55 73.40 22.70 3.90 Sandy Loam 0.38 30-60 1.55 2.57 66.89 28.39 4.72 Sandy Loam 0.40 60-80 1.54 2.56 68.57 26.54 4.89 Sandy Loam 0.40 80-100 1.54 2.58 69.10 26.54 4.36 Sandy Loam 0.40 100-140 1.59 2.62 68.01 27.38 4.61 Sandy Loam 0.39 21
The picture can't be displayed. Experimental Program... Laboratory Experiments Pressure Plate Experiment 22
Experimental Program... Laboratory Experiments Pressure Plate Experiment Suction Water Content (%) Pressure (cm) Sample 1 Sample 2 Sample 3 Sample 4 Sample 5 0 35.70 39.97 38.13 38.97 39.10 300 17.90 24.43 19.20 19.94 19.41 375 17.50 21.71 18.87 18.84 19.35 850 13.80 16.72 13.56 14.78 14.00 1900 7.90 8.56 7.55 8.92 10.68 5000 6.80 6.70 6.12 6.86 7.71 7000 7.00 7.13 6.07 7.11 7.84 9000 6.90 5.86 5.61 5.97 6.49 12000 6.90 7.15 6.56 6.58 7.95 moisture content (-) 0.5 0.4 0.3 0.2 0.1 0 Observed (0-30 cm) van Genuchten (0-30 cm) Observed (30-60) cm van Genuchten (30-60) Observed(60-80) cm van Genuchten(60-80 cm) Observed(80-100) cm van Genuchten (80-100 cm) Observed (100-140 cm) van Genuchten100-140 cm 0 5000 10000 15000 Suction pressure head(-ψ, cm) 23
Experimental Program... Laboratory Experiments Pressure Plate Experiment Soil Hydraulic Parameters Depth (cm) θ r θ s α (1/cm) n R 2 0-30 0.046 0.357 0.016 1.493 0.9751 30-60 0.056 0.399 0.006 1.741 0.9827 60-80 0.041 0.381 0.013 1.545 0.9732 80-100 0.018 0.390 0.022 1.366 0.9750 100-140 0.053 0.391 0.018 1.473 0.9855 Average 0.043 0.384 0.015 1.523 0.9783 Standard deviation 0.015 0.016 0.006 0.138 0.00544 24
Study site and Field experimental set up 25
Study site and experimental set up 26
Field preparation and Lysimeters 27
Transplanting paddy and sowing Berseem crops 28
Crop Growth stages and Irrigation 29
Crop Growth stages and Harvest 30
Crop Parameters Crop Parameters including Root depth, crop height and leaf area index (LAI) were monitored in each of the crop seasons. Root depth and crop Height were monitored using simple tape measurement for randomly selected crops and locations and the measurement values were averaged. LAI was monitored using L-80 ceptometer (leaf area monitoring device in field). 31
Root Depth 60 Root Depth(cm) 50 40 30 20 10 0 paddy rice (season-1) paddy rice (season-2) berseem fodder(season-1) berseem fodder(season- 2) 0 20 40 60 80 100 120 140 160 Number of days after transplanting/sowing) 32
Crop Height 1.6 Crop height (m) 1.2 0.8 0.4 0 Paddy Rice (Season-1) Paddy rice (sesaon-2) 0 20 40 60 80 100 Number of days after transplanting Crop height (m) 0.6 0.4 0.2 0 Berseem fodder(season-1) Berseem fodder (sesaon-2) 0 20 40 60 80 100 120 140 Number of days after sowing 33
Leaf Area Index LAI (m 2 /m 2 ) 8 7 6 5 4 3 2 1 0 Paddy Rice (Season-1) Paddy rice (sesaon-2) 0 20 40 60 80 100 Number of days after transplanting LAI(m 2 /m 2 ) 6 5 4 3 2 1 0 Berseem fodder(season-1) Berseem fodder (sesaon-2) 0 50 100 150 Number of days after sowing 34
Deep percolation Monitoring Using Field Lysimeters 35
Soil Moisture Monitoring Using Profile Probe (PR2/6) 36
Saturated soil Hydraulic Conductivity Using Guelph Permeameter 37
Crop Yield 38
Climatic Data The climatic data needed for the current study has been obtained from the nearby stations (800 m distance from the experimental station). The climatic variables are: Rainfall, Maximum and Minimum Temperature, Wind velocity, Relative Humidity and Sunshine hours all for daily time step. NIH and Department of Hydrology. 39
Results The Water Balance Model has a bit modified and used DP i nl = ( θ θ j= 1 i 1 i ) j where DP [L]= Deep percolation of water moving below the root zone; θ= is the volumetric soil moisture content (%); P[L] = rainfall; I [L]= applied irrigation; ET a [L]= actual evapotranspiration; R [L]= surface runoff, i and i-1 are, respectively, the current and previous time steps; j is an index for root zone layer and nl is the number layers. + P i + I i ET ai R i 40
Deep percolation (mm) 52 42 32 22 12 2-8 -18 Measured DP Computed DP -28 07/12/2013 06/01/2014 05/02/2014 07/03/2014 06/04/2014 06/05/2014 (a) Growing dates Deep percolation (mm) 54 44 34 24 14 4-6 Measured DP Computed DP -16 12/11/2014 12/12/2014 11/01/2015 10/02/2015 12/03/2015 11/04/2015 (b) Growing dates Fig. 6. Computed and measured deep percolation on daily time step in lysimetre 1 in berseem season 1 (a) and 1 in berseem season 2 (b) 41
Deep percolation (mm) (a) 120 100 80 60 40 20 0 27/12/2013 26/01/2014 25/02/2014 27/03/2014 26/04/2014 Growing dates Measured DP Computed DP Deep percolation (mm) (b) 65 55 45 35 25 15 5 Measured DP Computed DP -5 25/11/2014 25/12/2014 24/01/2015 23/02/2015 25/03/2015 Growing dates Fig. Computed and measured deep percolation with lumped time steps in lysimetre 1 in berseem season 1 (a) and in berseem season 2 (b) 42
In general: The performance of the simple water balance model is poor for daily time step while it performs well on the longer time step (lumped time step)as depicted in the above figures. 43
The Physically based model results: Calibration Deep percolation (mm) (a) 140 120 100 80 60 40 20 Measured DP Computed DP 0 20/07/2013 09/08/2013 29/08/2013 18/09/2013 08/10/2013 28/10/2013 Growing dates 50 Measured DP 40 Computed DP Deep percolation (mm) 30 20 10 0 10/12/2013 14/01/2014 18/02/2014 25/03/2014 29/04/2014 (b) Growing dates Measured and model predicted deep percolation in lysimetre 1 for rice season 1 (a) berseem season 1(b) 44
The physically based model performs well on daily as well as lumped time steps. Although, both models perform well on lumped time steps (weekly bases in this study), the physically based model performs superior than the simple water balance model. However, the benefit is compensated for large input data requirement in the case of physically based model. 45
Soil moisture content To verify the efficacy of the physically based model, computed soil moisture contents (obtained after calibration of the model using deep percolation data) and observed soil moisture contents were compared. The comparison yielded good results, although some discrepancy in estimating SWC by the model has been observed. 46
Water Productivity The water productivity (water use efficiency) of the crop is determined to evaluate the effect of water saving on crop yield. It can be expressed by following equations (Li et al. 2014; Sudhir-Yadav et al. 2011; Michael 2005): WWWW EEEEaa = YY EEEE aa WWWW II = YY II WWWW II+PP = YY II + PP where, WP ETa = water productivity based on evapotranspiration (Kg.m -3 ) Y = actual crop yield (Kg) ET a = actual evapotranspiration (m 3 ) WP I = water productivity based on irrigation input (Kg.m -3 ) I = irrigation input (m 3 ) WP I+P = water productivity based on total water input (Kg.m -3 )
Crop yield and water productivity indices for paddy (grain yield) Paddy season 1 Plot ID A11 A12 A13 A14 L2 L1 Average Yield (kg/ha) 4140.0 4140.0 4030.0 4860.0 3250.0 3540.0 ET a (mm) 408.64 411.72 411.72 411.72 410.9 410.99 I (mm) 2418.8 2418.8 2418.8 2418.8 2418.8 2418.80 I+P(mm) 3087.1 3087.1 3087.1 3087.1 3087.1 3087.10 WP ETa (Kg/m 3 ) 1.01 1.01 0.98 1.18 0.79 0.86 WP I (Kg/m 3 ) 0.17 0.17 0.17 0.20 0.13 0.15 WP (I+P) (Kg/m 3 ) 0.13 0.13 0.13 0.16 0.11 0.11 Paddy season 2 Yield (kg/ha) 3036.44 2666.67 2603.00 4700.0 2695.0 2688.17 ET a (mm) 430.87 430.91 431.05 430.62 414.04 431.05 I (mm) 643.1 639 855 644 851.00 630.00 I+P(mm) 1176 1171.9 1387.9 1176.9 1383.9 1162.9 WP ETa (Kg/m 3 ) 0.70 0.62 0.60 1.09 0.65 0.62 WP I (Kg/m 3 ) 0.47 0.42 0.30 0.73 0.32 0.43 WP (I+P) (Kg/m 3 ) 0.26 0.23 0.19 0.40 0.19 0.23 Yield decrease (%) 26.66 35.59 35.41 3.29 17.08 24.06 48
Crop yield and water productivity indices for berseem (green forage) Berseem season 1 Plot ID/Lysimeter A11 A12 A13 A14 L2 L1 Average Yield (kg/ha) 48900 53800 55500 51700 41200 37200 ET a (mm) 341.38 342.36 342.36 342.36 341.38 342.36 I (mm) 520.00 520.00 520.00 520.00 520.00 520.00 I+P(mm) 745.8 745.8 745.8 745.8 745.8 745.8 WP ETa (Kg/m 3 ) 14.32 15.71 16.21 15.10 12.07 10.87 WP I (Kg/m 3 ) 9.40 10.35 10.67 9.94 7.92 7.15 WP (I+P) (Kg/m 3 ) 6.56 7.21 7.44 6.93 5.52 4.99 Berseem season 2 Yield (kg/ha) 40641.2 49660.2 35944.2 40904.0 27333.0 27893.0 ET a (mm) 162.81 162.81 162.81 162.81 162.81 162.81 I (mm) 175.10 164.50 127.00 127.00 91.90 63.10 I+P(mm) 395.90 385.30 347.80 347.8 312.70 283.90 WP ETa (Kg/m 3 ) 24.96 30.50 22.08 25.12 16.79 17.13 WP I (Kg/m 3 ) 23.21 30.19 28.30 32.21 29.74 44.20 3 49
Conclusions Deep percolation computed using the water balance model on daily time step do not agree with field observed deep percolation for both crop seasons and lysimeters. However, the model predicts deep percolation very well on lumped time steps. Therefore, accurate estimation of deep percolation can be made on lumped time steps using simple water balance model. Physically based model, unlike the water balance model, predicts the deep percolation very well on daily time step. Both models predict DP very well on lumped time steps. 50
Conclusions The amount of deep percolation in both crop seasons is large. Deep percolation values ranging from 82 to 87% of input water has been lost through deep percolation in paddy season 1. In paddy season 2, deep percolation was 77-80% of the overall input water. In berseem season 1, the field observed deep percolation was 62-67% of input water while it has been reduced to 42-52% of input water in the berseem season 2. Increasing input water increases DP. 51
Conclusions Locally constructed drainage type lysimeters are demonstrated to be robust enough in capturing deep percolation from the bottom of crop root zones. The lysimeters were responding well to the imposed irrigation and rainfall events in the growing seasons of paddy and berseem crops subjected to varying regimes of water application. The lysimeters were also depicted the phenomena of preferential flow transport, distinguished the difference between daily and nocturnal deep percolation values. 52
Conclusions Simulations using physically based model also showed a visible association between the observed and model simulated soil moisture content in the soil profile. The performance of the physically based model shows that the model performs comparatively better for wet season than the dry season. 53
Conclusions The values of saturated hydraulic conductivity near soil surface is large. This would be attributed to root profile, the activities of soil fauna and soil cracking near the soil surface. It is possible to reduce deep percolation without the implementation of puddling practice in particularly in paddy fields and achieve large saving in input water by employing alternative irrigation scheduling strategy. 54
Conclusions Large saving in input water has been achieved with nominal yield decrease by employing alternative irrigation scheduling strategy during both crop seasons. Irrigation water saving on the other hand ranges from approximately 65% to 74% of the typical existing irrigation application for the rice crop in the region. On the other hand, in the berseem season input water saving ranging from 47% to 62% has been achieved. Irrigation water saving in the order of 66% to 88% of the conventional approach in berseem has been attained. 55
Conclusions There was yield reduction due to the reduced application of irrigation. However, the water productivity has been increased. Therefore, it can also be concluded that increased water productivity in a given field can be realized by altering an irrigation scheduling strategy. 56
Scope of Future Work The current study was limited to few number of experimental trials. Large trial experiments may be needed to asses an optimum irrigation scheduling strategy which would reduce DP and provide optimum water productivity for a given cropping condition. The current work may be extended to other soil types for the rice, berseem and other cropping conditions to quantify and investigate deep percolation characteristics. 57
Scope of Future Work In the current study due consideration is given for single porosity model, assuming matrix flow conditions. Future works may also consider macropore flow conditions. The effect of other variables (crop variety, agronomic conditions, climatic conditions etc...) on crop yield may be investigated in future. The locally constructed drainage type lysimeters play an important role in monitoring deep percolation. These types of lysimeters may be constructed elsewhere to study groundwater recharge, solute transport etc. 58
Thank You ध व द