HYDROLOGICAL RESPONSE OF UNSATURATED ZONE UPTO WATER TABLE

Abstract

The rainfall and applied irrigation exoite the unsatura ted zone, extending from ground surface upto the water table,at the ground surface. The hydrological response to this excitation comprises amongst others, the return flow from the applied irri gation and rainfall, change in soil moisture storage and ponding of water. The return flow is an important component of groundwater resource of an unconfined aquifer. The soil moisture status in the root zone and the ponding of water has an important bearing on the evapotranspiration of vegetation. The evapotranspiration in **§* of agricultural crops determines the necessity of supple menting rainfall by irrigation, for maintaining a predefined mois^re level in part of/entire root zone or for maintaining a range of ponded depth of water. Thus, a quantitative estimate of the response of the unsaturated zone is a prerequisite for carrying out ground water, crop water requirement and other related studies. The response is governed by the unsteady state moisture flow in the zone. Current practice, of simulating the flow process, is mostly based upon soil moisture accounting (SMA) models considering the entire zone as a unit. However, these models suffer from many restrictive assumptions. Prominent amongst them is the assumption of existence of a threshold moisture content (termed as field capacity), below and at which there occurs no moisture movement and above which the excess moisture is drained in the basic accounting period, irrespective of the soil drainability. This could lead to discrepancies in time distribution as well as Periodical totals of return flow. Many of these assumptions can be iv eliminated by solving the governing differential equation of the unsteady state flow of moisture in the zone. This equation is known as Richards equation. Since, solution of the equation can provide time and space distribution of 'the response, the models governed by this equation would be distributed models. The present work is an attempt, to develop a one dimen sional (vertical) distributed numerical model, to simulate the unsteady flow of water in the unsaturated zone involving evapo transpiration. The mode3. is based upon colution of the Richards equation by Crank-Nicolson finite difference scheme. An algorithm has been developed for identification and assignment of the upper boundary condition (ground surface boundary condition). However, the overland flow is not simulated. The lower boundary oondition is assigned accounting for the time variant position of water table. Further, a piecewise continuous functional relation for capillary suction head versus moisture content and an anpirical oriteria for specifying variable time step of simulation have been developed. Caloulations of the model are performed with the assistance of a digital computer. The computer code has been written in FORTRAN IV. The moisture profiles simulated by the model compare well with those given by the Philips quasi-analytical solution (Philip 1969), for a soil(yolo light cloy) under identical conditions. Further, the model has been operated to simulate moisture profiles of a layered soil under field situation. The simulated moisture profiles compare will with the observed profiles. Statistical evaluation, of the model simulation, in respect of moisture profiles, by calculating coefficient of correlation, F raid t statistics indi cates a satisfactory performance of the model. V The model has been operated to schedule irrigation for a few soil-crop conditions, under daily rainfall series of a normal rainfall year reported from a local rainguage station. The soil crop conditions considered werej rice-wheat cropping on clay as well as on loam and sugarcane on loam. Irrigation criterion con sidered for wheat and sugarcane was 50 percent allowable average moisture depletion in the entire root zone. For rice two different criterion have been considered. These are:no allowable average moisture depletion in the entire root zone (upland cultivation), and requirement of maintaining a minimum ponding of 50 mm (low land or submergence cultivation). Time distributions of the return flow (from the rainfall and the scheduled irrigation) have been worked out, for the rice-wheat cropping on clay as well as on loam. The model scheduled irrigation totals in case of upland rice, wheat and sugarcane, are generally lower than the generally existing local practice. The major reason. for this deviatior was suspected to be the farmers'practice of irrigating by 'feeling' the moisture depletion in the upper part of the root zone only, where the moisture depletion would be relatively faster. In order to verify this argument the model was re-operated for the wheat cropping on loam,with a modified irrigation criterion of no allowable average moisture depletion in the top 30 cms (usual tillage depth) of the root zone . The irrigation so scheduled was quite close to the generally existing local practice. The model scheduled total irrigation in case of low land rice cultivation has been in the reported range of the existing practices in India. Annual return flows have worked out to 71.61 percent and 50.67 percent, of the rainfall and applied irrigation, incase of rice-wheat cropping on clay and on loam respectively. It has been noticed vi that due to a time lag (between occurence of input at ground surface and occurence or return flow at the water table); the return flow in certain time periods (months) are dispropor tionate to the corresponding inputs. The current practice of quantifying field capacity by adopting the mois^re content corresponding to 0.1 to 0.5 bar tension may not always be consistent with it s hydraulic im plication in the SMA models. So a method has been proposed to quantify field capacity as a flow parameter, to be more objective. This method is more suitable for coarser soils. A SMA model has been operated to route infiltration through the unsaturated zone. Field capacity in this model is quantified as per the proposed method. The daily infiltration series generated while calculating return flows (for the ricewheat cropping on clay andon!oan)by the distributed model, have been routed through the unsaturated zone, by this model. The SMA model over estimated the return flow rates during the early (rice crop) period, in comparison to the return flow rates given by the distributed model. Subsequently, these return flow rates were lower. This situation is more pronounced with the clay soil. This descrepancy is due to the fact that,the SMA model doesn't account for the time lag in occurrence of the return flow. As a result of such an underestimation and over estimation of return flow during different simulation periods the errors in the estima tes of seasonal totals got compensated to some extent. Thus, the seasonal totals of return flows computed by the SMA model tended to match with the corresponding totals arrived at by the distri buted model