Integrating micro-flood irrigation with in-field rainwater harvesting

[摘要] The mam aim of the study was to integrate micro-flood irrigation (MFI) with in-fieldrainwater harvesting (IRWH). The MFI is a short furrow irrigation system that relies on smallinflow rates to mitigate the effect of dry spells in crop fields. The IRWH is an in situ basedrainwater harvesting technique that harvests rainfall in the form of runoff between crop rowsand then concentrates it in the basin area. Given the increased rainfall variability andevaporation (Ev) in the semi arid areas of the central Free State Province of South Africa, themerging of these two technologies is hypothesized to be able to stabilize soil water storageduring rainfall and dry spell periods in areas with access to limited irrigation water.The developments in the study were divided into three phases. The first phase dealt withcharacterization of pedological and hydraulic properties of the soils earmarked for IRWH atthe University of the Free State, 20 Km, south of Bloemfontein. These soils were representedby the Tukulu, Sepane and Swartland soil types with the first two forms also referred asCutanic Luvisols and the latter as Cutanic Cambisols of the Reference Soil Group. Thesesoils were similar only in the orthic A- horizon. The Tukulu had developed structure only inthe prismatic C-horizon and for the Sepane it was in the pedocutanic B- and prismatic Chorizons.The Swartland had a cambic structure in the pedocutanic B-horizon. Correspondinghydraulic properties, soil water characteristic curve (SWCC) and hydraulic conductivity forsaturated (Ks) and unsaturated conditions (K-8) were determined using in situ and laboratoryprocedures for internal drainage (ID) and evaporation (Ev) conditions. Parametric modelswere used to describe SWCC and to predict K-8 relationships. Model descriptions of SWCCwere satisfactory. Predictions of K-8 were only accurate at near saturation, but HYDRUS-IDoptimization program had better predictions. Matric suction gradients corresponding to thedraining soil profile were found to fall within the matric suction range of 0 to -10 kPa.Drainage rate of 0.001 mm hour corresponded to drainage upper limit (DUL) and deepdrainage (DD) losses proportional to 1 % of annual rainfall over the fallow period. TheTukulu, Sepane and Swartland soil types had respectively total DD losses of 21, 20 and 52mm and evaporation losses of 43, 51 and 70 mm. The Ks corresponding to the C-horizons ofthese soils was 9.6, 1 and 77 mm hour. During ID and Ev the K-8 functions especially forhorizons with a clay content range of 26 to 48 % dropped by several orders of magnitudes,while SWC changed with a narrow margin. At the evaporating surface matric suction ofmagni tude greater than -1500 kPa were approximated.The second phase compared four inflow rates (20,40, 80 and 160 L min-I) based on surfaceand subsurface irrigation characteristics carried out on the Tukulu soil due its low DD and Evlosses. A single irrigation on a 90 m closed ended furrow and measurements taken at every10 m furrow distance for advance and opportunity times, stream flow depth, and SWC beforeand after the irrigation. Infiltrated depths predictions from HYDRUS-2D software weresatisfactory from all inflow rates. Distribution uniformity (DU) was higher (�?0.89) at 30 mfurrow distance from all inflow rates and the smaller inflow rate much easier to handle.Vertical redistribution was characterized at each of the 10 m furrow distance covered with a 2m x 2 m polythene sheet to prevent Ev. Over the 455 hours of redistribution agreementbetween measured and predicted SWC from HYDRUS-2D software varied with depth andfurrow length. Low vertical redistribution (Vz) from all inflow rates was attributed to therestrictive prismatic C-horizon. Higher rates of Vz were observed within the 0-600 mmprofile domain for the small inflow rates and at 0-850 mm for the large inflow rates.The last phase dealt with the integration of MFI with IRWH, carried out on a 3 x 3 split plotfactorial with four blocks in a complete randomized design experiment. Each plot had five 30m long furrows and a pair of neutron access tubes installed in each plot at the centre of thebasin and runoff area. The main treatments were runoff strip width (RSW; 1 m, 2 m and 3 m)and water regime (WR): dryland (DL), supplemental (SPI) and full irrigation (FI). No till andbasin tillage was used to prepare the RSW and the 1 m standard basin area (BA). The BA wasfurther smoothed with a ridger for uniform distribution of the advance stream flow. A 120day maturing maize variety was used. A record of rainfall and SWC was kept. The 40 L min-Iinflow rate for 15 minutes irrigation times on a fixed schedule for full and supplementalirrigation, provided by the BEW AB+ irrigation software was used. A soil water balance(SWB) procedure was developed to evaluate the effect of the RSW and WR treatments on thegains and losses in soil water storage. Evapo-transpiration (ET) was partitioned into Ev andtranspiration (T) by a ~-parameter based on plant canopy area. Findings showed that SWBcomponents were affected by the main effect from the RSW and WR. The 1 m RSW had thetotal biomass and grain yields that were respectively 21 % and 45 % higher than the 2 mRSW, and 35 % and 89 % higher than the 3 m RSW. Total biomass and grain yields from fulland supplemental irrigation were 200 % and 76 % higher than the DL. Though tested for asingle season the combination of 1 m RSW and full irrigation produced optimum crop yieldsand WUE for the newly merged MFI-IRWH water management system and is ready to beused by small scale farmers who have access to irrigation water.

[发布机构] University of the Free State