Soil water dynamics and response of cowpea to water availability under moisture irrigation.

Abstract

Increasing population, urbanization and industrialization has put pressure on the irrigation sub-sector to produce more yield using less water i.e. improving crop water productivity (WP). This can be achieved through the adoption of efficient irrigation systems such as micro-irrigation. Moistube irrigation (MTI) is a relatively new technology like subsurface drip irrigation (SDI) but with a semi-permeable membrane whose nanopores emit water in response to applied pressure and soil water potential. Being a new technology, there is little information regarding its hydraulic characteristics and soil water distribution which are necessary for its design, operation and management. Furthermore, the response of crops under a variety of soils and environmental conditions under MTI has not been covered extensively. Therefore, this study aimed at determining the hydraulic and clogging characteristics of MTI. The effect of soil texture on the soil water dynamics of MTI was also determined. Finally, the response of cowpea, an important but neglected African indigenous legume, to varying water regimes under MTI was also determined. This study was based on the hypothesis that cowpea responds favourably to water regimes under MTI. The study was accomplished through laboratory, field experiments and agro-hydrological models. AquaCrop and HYDRUS 2D/3D were chosen for this study due to their reliability in predicting crop yield responses to water availability and soil water dynamics respectively. The laboratory experiments were conducted in soil bins to determine the soil water dynamics of MTI under sandy clay and loamy sand soils which were used to calibrate the HYDRUS 2D/3D model. The hydraulic characteristics were determined at a pressure of between 10 kPa and 100 kPa while the effect of suspended and dissolved solids was determined under a pressure of 20 kPa and 30 kPa. The field experiments consisted of glasshouse and tunnels to examine the response of cowpea to full and deficit irrigation of MTI with SDI as the control. The results were used to parameterise and validate the AquaCrop model. Finally, HYDRUS 2D/3D and AquaCrop were coupled to draw into the strengths of the individual models and used to simulate the water use of cowpea under MTI in two agro-ecological zones in South Africa. The results showed that the discharge – pressure relationship of Moistube followed linear and power functions. It was also established that suspended solids had severe clogging effect than dissolved solids. In the soil bin experiment, simulated water contents closely matched (R2 ≥ 0.70 and RMSE ≤ 0.045 cm3 cm-3) the observed values in all the points considered for the two soil textures. The model slightly under-estimated or over-estimated the soil water content with percent bias less than 15.6%. There was no significant difference (p > 0.05) between the soil water distribution in lateral and downward direction for both sandy clay loam soil and loamy sand. However, the soil water content upward of the Moistube placement depth was significantly lower (p < 0.05) than both the lateral and downward soil water contents in loamy sand. The soil water dynamics under MTI while incorporating the root water uptake indicated that there was no significant difference between the root water uptake in SDI and MTI (p > 0.05). Water loss through drainage was significantly higher (p < 0.05) under SDI than MTI in loam while it was negligible in clay for both irrigation types. Drainage increased with increased Moistube placement depth. The interaction between the distribution of root water uptake and

the soil water distribution indicated that a suitable placement depth for cowpea under MTI was 15 cm in loam and 20 cm in clay. There were no significant differences (p > 0.05) in the yield response of cowpea between MTI and SDI but the latter performed better under deficit irrigation conditions. AquaCrop model was parameterized and tested successfully under full and deficit irrigation. The results indicated the model simulated the canopy cover (CC) very well with R2 ≥ 0.85, RMSE ≤ 24.5%, EF ≥ 0.45, and d ≥ 0.87. The simulated water content closely matched the observed with R2 ≥ 0.61, RMSE ≤ 11.3 mm, EF ≥ 0.51, and d ≥ 0.86 indicating that the model reasonably captured the soil water dynamics. Generally, yield and biomass were simulated satisfactorily by the model with R2 of 0.84 and 0.88, and RMSE of 282 kg ha -1 and 1307 kg ha -1, respectively, during parameterisation. Similarly, during model testing the model performance was very good with R2 of 0.96 and 0.99, and RMSE of 165 kg ha -1 and 798 kg ha -1 for yield and biomass, respectively. The highest WP was achieved under 70% ETc (crop water requirement) and 40% ETc for yield and biomass, respectively. Having successfully calibrated and tested the HYDRUS 2D/3D and AquaCrop models, the two were used symbiotically to simulate the water use of cowpea in two environments characterized by clay and sandy soils. The crop characteristics were obtained using AquaCrop while HYDRUS 2D/3D was used to generate optimum irrigation schedules and the soil water balance. Thereafter, the water use and yield of cowpea was determined. The average grain yield and biomass were 2600 kg ha-1 and 10000 kg ha-1, respectively, with the difference between the two sites being less than 5% under both SDI and MTI. The water use and WP varied from 315 mm to 360 mm and 0.67 to 1.02 kg m-3, respectively, under the two irrigation types at the two sites considered. The WP was higher under SDI than MTI, but the differences were less than 10%. This showed that cowpea responded similarly under MTI and SDI. Further research is needed on the determination of the clogging characteristics due to fertigation. Finally, more field experiments under other environmental conditions need to be carried out to validate the results of this study.