Integrating satellite-based evapotranspiration with simulation models for irrigation management at the scheme level

Improvements in irrigation management are urgently needed in regions where water resources for irrigation are being depleted. This paper combines a water balance model with satellite-based remote-sensing estimates of evapotranspiration (ET) to provide accurate irrigation scheduling guidelines for individual fields. The satellite-derived ET was used in the daily soil water balance model to improve accuracy of field-by-field ET demands and subsequent field-scale irrigation schedules. The combination of satellite-based ET with daily soil water balance incorporates the advantages of satellite remote-sensing and daily calculation time steps, namely, high spatial resolution and high temporal resolution. The procedure was applied to Genil–Cabra Irrigation Scheme of Spain, where irrigation water supply is often limited by regional drought. Compared with traditional applications of water balance models (i.e. without the satellite-based ET), the combined procedure provided significant improvements in irrigation schedules for both the average condition and when considering field-to-field variability. A 24% reduction in application of water was estimated for cotton if the improved irrigation schedules were followed. Irrigation efficiency calculated using satellite-based ET and actual applied irrigation water helped to identify specific agricultural fields experiencing problems in water management, as well as to estimate general irrigation efficiencies of the scheme by irrigation and crop type. Estimation of field irrigation efficiency ranged from 0.72 for cotton to 0.90 for sugar beet.     

Assessing water consumption of banana: traditional versus modelling approach

The water consumption, expressed as ‘actual evapotranspiration’ (ET_a), plays an important role in understanding the soil–water–atmosphere–plant environment. This paper highlights two different methods to assess the water consumption of banana, a traditional and modelling approach, respectively.The first method to assess ET_a was based on the water balance equation and thus included field measurements. The second method involved the simulation model soil water atmosphere plant (SWAP). The results showed that both methods led to comparative results.This study also aimed at evaluating the performance of the SWAP-model under very different circumstances as for which it has been used previously. The uncertainty about the parameters of the SWAP model remained rather high. The most important drawback of SWAP is the enormous amount of data that is required. Since these data were not always available, this resulted in a large amount of parameters to be estimated and involved in the sensitivity analyses and calibration process. The final simulation results showed several inconsistencies. It seemed that, for successful predictions, more data were needed to describe the bottom boundary condition (BBC) of the considered soil profile in greater detail, especially when there is influence of groundwater. This BBC has indeed a great impact on the simulation results. SWAP offers several ‘simplified’ approaches concerning the BBC in case no data are available. However, in all cases SWAP seemed to overestimate the moisture content.     

Evapotranspiration data assimilation with genetic algorithms and SWAP model for on-demand irrigation

Evapotranspiration (ET) is one of the indicators of water use efficiency. Periodic information of ET based on remote sensing is useful for an on-demand irrigation (ODI) management. The main objective of this paper was to develop an ET data assimilation scheme to optimize the parameters of an agro-hydrology model for ODI scheduling. The soil, water, atmosphere, and plant (SWAP) simulation model has been utilized for this purpose. We computed remote sensing-based ET for a wheat field in the Sirsa Irrigation Circle, Haryana, in India using 18 cloud-free moderate resolution imaging spectroradiometer images taken between December 2001 and April 2002. The surface energy balance algorithm for land (SEBAL) was used for this purpose. Because ET estimates from SEBAL provide information on the surface soil moisture state, they were treated as observations to estimate unknown parameters of the SWAP model via a stochastic data assimilation (genetic algorithm) approach. The SWAP parameters were optimized by minimizing the residuals between SEBAL and SWAP model-based ET values. The optimized parameters were used as input to SWAP to estimate soil water balance for ODI scheduling. The results showed that the selected parameters (i.e. sowing, harvesting, and irrigation scheduling dates) were successfully estimated with the data assimilation methodology. The SWAP model produced reasonable states of water balance by assimilating ET observations. The root mean square of error was 0.755 and 2.132 cm³/cm³ for 0–15 and 15–30 cm soil depths the same layers, respectively. With optimized parameters for ODI, SWAP predicted higher yield and water use efficiency than traditional farmer’s irrigation criteria. The data assimilation methodology produced can be considered as an operational tool at the field scale to schedule irrigation or predict irrigation requirements from remote sensing-based ET.     

Performance assessment of an irrigation scheme using indicators determined with remote sensing techniques

In this work, remote sensing-based assessments of actual evapotranspiration using METRIC integrated with a water balance model provided accurate estimates of irrigation performance. This new methodology was applied and tested in the Genil–Cabra Irrigation Scheme located in southern Spain during the 2004–2005 irrigation season. The performance indicators used, the annual relative irrigation supply (ARIS) and the irrigation water productivity (IWP), required ET input data which were calculated using either METRIC or standard FAO methodology. The new procedure that used METRIC detected overirrigation (ARIS of 1.27) in situations where the ARIS calculated with the standard FAO methodology indicated near-optimal irrigation (ARIS of 0.98). Additionally, the proposed methodology allows the estimation of the volume of applied water at the field scale. Comparisons between the ARIS and IWP values obtained from actual applied water records against those calculated with the new methodology resulted in good agreement. It is concluded that the integration of the METRIC method to calculate actual ET with a water balance model allowed the determination of performance indicators in an irrigation scheme in a reliable and accurate fashion, requiring only very limited information at the field level.     

Application of the optimization method for estimating infiltration characteristics in furrow irrigation and its comparison with other methods

The infiltration characteristics of a soil vary spatially and temporally, and due to this the available methods for estimating the characteristics in furrow irrigation are either not suitable or have restrictions for their field use. An optimization method based on the volume balance approach, originally developed for estimating infiltration parameters in border irrigation, and using multiple observations of arrival time of the wetting front was modified for furrow irrigation. The method was applied to 13 irrigation events on furrows monitored on a farm in northern New South Wales, Australia. The soil type at the experimental site has a high clay content (up to 67%) and develops cracks when dry. In addition to the optimization method, one-point and two-point methods using observations of arrival time reported in the literature were also used. The accuracy of different methods was evaluated by comparing the calculated total volume of water infiltrated into the furrow with that observed in the field. The optimization method was the most accurate and the one-point and the two-point were the least accurate among three methods considered in the present study. A possible explanation for a poor performance of the one-point and two-point methods might be related to the assumptions made in the derivation of the methods and the unsuitability of the Philip and the Kostiakov infiltration equations used for the field condition in the present study.     

Combining remote sensing-simulation modeling and genetic algorithm optimization to explore water management options in irrigated agriculture

We present an innovative approach to explore water management options in irrigated agriculture considering the constraints of water availability and the heterogeneity of irrigation system properties. The method is two-folds: (i) system characterization using a stochastic data assimilation procedure where the irrigation system properties and operational management practices are estimated using remote sensing (RS) data; and (ii) water management optimization where we explored water management options under various levels of water availability. We set up a soil–water–atmosphere–plant model (SWAP) in a deterministic–stochastic mode for regional modeling. The distributed data, e.g. sowing dates, irrigation practices, soil properties, depth to groundwater and water quality, required as inputs for the regional modeling were estimated by minimizing the residuals between the distributions of field-scale evapotranspiration (ET) simulated by the regional application of SWAP, and by surface energy balance algorithm for land (SEBAL) using two Landsat7 ETM+ images. The derived distributed data were used as inputs in exploring water management options. Genetic algorithm was used in data assimilation and water management optimizations. The case study was conducted in Bata minor (lateral canal), Kaithal, Haryana, India during 2000–2001 rabi (dry) season. Our results showed that under limited water condition, regional wheat yield could improve further if water and crop management practices are considered simultaneously and not independently. Adjusting sowing dates and their distribution in the irrigated area could improve the regional yield, which also complements the practice of deficit irrigation when water availability is largely a constraint. This result was also found in agreement with the scenario that water is non-limited with the exception that the farmers have more degrees of freedom in their agricultural activities. An improvement of the regional yield to 8.5% is expected under the current scenario.     

Evaluation of irrigation schemes for sugarcane in Sindh,Pakistan, using SWAP93

A computer simulation model, SWAP93, was used to simulate the soil water balance of sugarcane (Saccharum officinarum L.) over a period of 6 years, in order to develop an efficient irrigation scheduling scheme for Sindh, Pakistan. Given the limitations and inflexibility of the existing warabandi irrigation system, which does not allow on-demand irrigation, only irrigation depth and irrigation interval were varied in order to assess the best irrigation depth/interval combination for sugarcane production. Twelve irrigation treatments were simulated. These treatments were four irrigation amounts (900, 1200, 1650 and 1800 mm) and three irrigation frequencies (7, 10 and 15 days). Three seasons with rainfall totaling less than 20 mm were compared with three seasons of over 200 mm rainfall. Two approaches were used in assessing the irrigation schemes: yield parameters and water management response indicators. Treatment parameters (e.g. irrigation amounts, weather conditions, soil characteristics, etc.) served as input for SWAP93, actual transpiration was calculated and then used in a crop water production function to predict yield and water use efficiency. Additionally, water management response indicators were derived from model outputs, and used to assess the impact of the schemes on soil salinity and water logging. Both these indicators and the yield and water use efficiency indicated that a seasonal total of 1650 mm, applied at a 15-day interval was the best irrigation scheme for the region.     

A mathematical model for simulating water balances in cropped sandy soil with conventional flood irrigation applied

In this paper, a model that integrates various complex model components for the purposes of water balance modeling throughout crop development in arid inland region under the conventional flood irrigation practiced is presented. These components are modules for calculating dynamic soil water content based Richard's equation, potential and actual evapotranspiration, and crop root water uptake. Soil water content in the active root zone and soil evaporation simulation obtained from the model were test using field data in 2003. The low values of MARE and high values of R² and PE in the active root zone of soil profile as well as daily soil evaporation indicated that the soil water balance simulation model presented in the paper can be used with reliable accuracy to simulate the components of water balance in cropped sandy soil under the conventional flood irrigation condition in arid inland regions. The model simulation on components of water balance using observed field data in 2004 indicated that large quantities – about 43% of irrigation water (amounting to 840 mm) – were consumed by deep percolation, only small (less than 41%) proportions of irrigation water used by the plants for transpiration. The current irrigation scheme is characterized by the unreasonable agricultural water management with the waste of water in the irrigational system in this region. The impact of irrigation scheduling on water balance presented in this paper showed that the reasonable irrigation scheme with more frequent irrigation and less amounts is more suitable for the irrigation of spring wheat in Heihe River basin, northwest China. Therefore, to establish a decision-making system for agricultural irrigation scheme and to utilize the limited water resources in this region have become an urgent problem that needs to be solved.     

Irrigation frequency and total water application with trickle and furrow systems

One of the main attributes of trickle irrigation is that it enables the available water to be utilised as efficiently as possible. This is achieved through the elimination of waste that is an inevitable consequence of normal irrigation practice.The water requirements of crops can be calculated to a reasonable degree of accuracy using formulae that have been developed to relate evapotranspiration to the evaporation from a U.S. Class A pan evaporimeter, with suitable adjustments for the particular crop and the degree of ground cover.The aims of the experiments described in this paper were:

1. (1) to compare the yield of field tomatoes under trickle irrigation and furrow irrigation over a range of irrigation frequencies and amounts that can reasonably be used for both methods;

2. (2) to compare the effects of irrigation frequency over a range of irrigation amounts that are typical for each irrigation method;

3. (3) to compare the effects of irrigation amounts over a range of irrigation frequencies that are typical for each irrigation method;

4. (4) to examine the interaction, if any, between irrigation frequency and amount.

The results of the trial showed that there was no difference in yield between trickle irrigation and furrow irrigation over a range of conditions for which they could legitimately be compared. At the same time it should be realized that the results for furrow irrigation on small plots may not adequately represent the normal farm practice for this method, where less efficient watering occurs due to long runs and unavoidable inequalities in water application.For both irrigation methods there was a consistent trend of decreasing yields with increasing irrigation frequency. Where departures from this trend occurred in unusual irrigation conditions (i.e. continuous irrigation and prolonged drought) the effect is explained by physiological causes that relate to the particular conditions. In such cases only, there was an interaction between irrigation frequency and amount.For each irrigation frequency the highest yields were obtained for the optimum amounts as calculated to match the potential evapotranspiration rate. Equal amounts of under-watering and over-watering resulted in approximately the same reductions in yields.The results from the trial indicate that the highest yields and greatest efficiency of water use occur at frequent intervals with an amount of applied water approximately equal to that calculated from simple formulae relating evapotranspiration to measured pan evaporation. Extrapolation of the results of particular trials to other crops, soils and climates should be treated with caution. However, the agreement of the results with those predicted from physiological and empirical data, and the fact that the results from other experiments also agree with such formulae, give cause for confidence in predicting the responses to irrigation timing and amounts over a wide range of conditions.     

浅埋条件下天然羊草群落生育期蒸散量的计算方法

该文以荷兰Wageningen农业大学等单位开发的土壤-水分-大气-作物系统模拟模型(SWAP模型)为基础,模拟计算浑善达克沙地地下水浅埋区天然羊草群落生育期蒸散量。为了确保模拟精度,首先在2003年野外试验以及室内试验所得的土壤水分数据基础上,利用2005年实测气象资料及作物指标,以2005年实测的土层剖面水分数据,对SWAP模型的参数进行率定,得到适合当地条件的模型参数;然后用2006年的数据对选定的模型参数进行检验和确定。结果表明:SWAP模型有良好的精度,可用于模拟地下水浅埋区蒸散量。     

基于CRAE模型的西北地区实际蒸散发量计算

应用我国西北地区常规气象资料对Morton互补相关(CRAE)模型进行了必要的校正,以适应其在西北地区的应用。结果表明,净长波辐射公式是影响模型计算精度的主要因素,且在校正中必须考虑高程的影响。校正后的模型可利用我国常规气象观测资料进行实际蒸散发量的计算,其精度可满足实际应用的要求,模型计算结果反映了西北地区水分蒸发的主要控制因素,具有一定的实际应用价值。     

Surface energy balance using satellite data for the water balance of a traditional irrigation—wetland system in SW Iran

A water balance of a large traditional irrigation area and a downstream adjoining wetland was determined using the surface energy balance approach (SEB), based on satellite data, to calculate the actual evaporation of both the irrigated area and the wetland at four different dates in a dry year and information of two additional images. The contribution of capillary flow by the shallow groundwater table was estimated by evaluating the actual evapotranspiration values of adjoining rangelands and non-inundated wetland areas. Those values were used to separate the total evapotranspiration into a soil moisture change component due to capillary rise, and into a component attributable to supply of river water. The only field data used for the estimated monthly water balance were air temperature, wind speed, and water inflow, since rainfall and outflow could be ignored in the year 2000. The results provided an insight for conditions of a drought year within the irrigated area, the distribution of water to irrigation and the wetland and showed the linkage between inundated wetland area and discharge.     

Semi-empirical approach for estimating actual evapotranspiration in Greece

Reliable estimation of actual evapotranspiration rates (ET) may be achieved if extensive information on the soil–plant–atmosphere system is available. Agricultural and irrigation engineers facing the problem of rational irrigation planning, rarely have at their disposal such information. Therefore, there is a demand for simpler approaches to estimate actual evapotranspiration. In this paper, a semi-empirical approach is proposed for estimating actual water losses from crops. It is assumed that the ratio of actual to maximum evapotranspiration (ET/ET_m) is an exponential function of the water content w in the root-zone, of the form: ET/ET_m=exp[c(w−w_fc)/(w−w_wp)], where c is a constant introduced to adjust the decrease of the ratio ET/ET_m according to what is observed for the climatic conditions, soils and crops of Greece, w_fc the water content at field capacity and w_wp denotes the water content at wilting point. Verification of the above approach for estimating actual evapotranspiration was achieved by comparing ET-values obtained by the soil moisture profile changes and the ET-values obtained by the above equation. Meteorological, crop and soil data required were collected from experimental fields of the Agricultural University of Athens (38°23′N, 23°6′E). The agreement of actual versus computed ET-values for three widely grown crops in Greece (cotton, wheat and maize) may be considered as satisfactory.     

Application of a simulation based optimization model for winter wheat irrigation scheduling in North China

Irrigation needs to be scheduled properly for winter wheat, the main food crop in North China where the water resources are limited. We optimized the irrigation timing of crops under limited water supply by integrating a soil water balance model, dated water production function with cumulative function of water sensitivity index, and a nonlinear search method. The optimization produced the optimal irrigation date series with the predetermined irrigation quota for each application, which aims to obtain higher crop yield with limited irrigation water and be convenient for irrigation management. This simulation–optimization model was used to investigate the irrigation scheduling of winter wheat in Xiaohe irrigation Area in North China. Results show that optimal irrigation date series, corresponding relative yield and relative evapotranspiration are all closely related to the irrigation quota and initial soil water conditions. For rich and medium initial soil water conditions in medium precipitation year, it takes four times of irrigation (60 mm each time) after greening in order to obtain higher crop yield. But it increases to five times for poor initial condition. With limited irrigation water, irrigation should generally be applied in the preferential sequence of early May or late April (in the jointing stage), then mid and late May (in the heading stage), and finally March (in the greening stage). Irrigation should be applied earlier with lower initial soil water storage. Higher irrigation quota increases the crop yield but tends to decrease the marginal value, especially when irrigation quota exceeds 180 mm. The study also indicates that the optimized relative yield is generally higher than that obtained in field experiment. Based on the optimization, we proposed to use the quadratic polynomial function to describe the frontier water production function, which shows the mathematical relationship between optimized relative yield and relative evapotranspiration.     

基于遥感的农业用水效率评价方法研究进展

遥感技术的发展为区域尺度蒸散发计算、作物分布识别及估产提供了一条有效途径,为基于遥感信息的灌区灌溉水利用效率及作物水分利用效率定量评价奠定了基础。回顾总结了遥感蒸散发模型、瞬时蒸散发升尺度方法、日蒸散发插值方法、作物分布识别方法及作物估产模型的研究进展,评述了遥感蒸散发及作物估产结果在灌区灌溉水利用效率及作物水分利用效率评价中的应用情况。提出了相关领域需要进一步研究的问题,包括适合非均匀下垫面特点且具有较强物理基础的灌区遥感蒸散发模型、日蒸散发插值中灌溉或降雨引起土壤含水量突变情况的处理、农田蒸散发中灌溉水有效消耗量的准确估算、能适应复杂种植结构并且适用于多年的作物分布遥感识别模型以及精度较高且可操作性强的遥感估产模型等。     

Representation of rainfed valley ricefields using a soil-water balance model

Soil-water conditions for ricefields located in valleys in micro-catchments are simulated using a daily soil-water balance model. The crop is primarily rainfed but there is also limited irrigation water. The simulation covers a complete year and includes features such as rainfall, irrigation releases, runoff from uplands, actual evaporation and evapotranspiration, percolation losses through the bed and bunds of the ricefield, standing water in the field and overflows from the ricefield. A specific location in Sri Lanka is selected to illustrate the approach. The impacts of different conditions are explored including alternative irrigation releases, increased losses through the bed and bunds of the ricefield and a lower overflow from the ricefield. Simulations indicate that ricefields which are towards the valley sides have an increased inflow due to runoff from adjacent uplands; this can lead to improved rice yields. However, reducing heights of the bunds to half the original value results in substantial overflows during periods of high rainfall while the number of days without submergence almost doubles. This uncomplicated model is consistent with the limited field data and information available; it provides a realistic representation of the important processes and indicates why poor crop yields often occur.     

Simulation of soil water in space and time using an agro-hydrological model and remote sensing techniques

Complete knowledge of all components of the water balance is essential to optimize water use in irrigated agriculture. However, most water balance components are very difficult to measure in terms of the required time interval and due to the complexity of the processes. An unsaturated zone model is a useful tool for predicting the effects of agricultural management on crop water use and can be used to optimize agricultural practices in view of minimizing the agricultural water use. For the irrigated areas in Minqin County of northwest China, the physically based one-dimensional agro-hydrological model SWAP (Soil, Water, Atmosphere and Plant) for water movement and crop growth was applied to reveal all the components of the water balance at multiple sites. This model has a varying level of abstraction referring to simulated processes in time and space. A combination of field, meteorological and aerial data was used as input to the model. Inverse modeling of evapotranspiration (ET) fluxes was followed to calibrate the soil hydraulic functions by using the parameter estimation package PEST. Surface Energy Balance System (SEBS) was used to estimate actual ET fluxes from NOAA AVHRR satellite images. Simulations were carried out for 15 different sites in Minqin County by using wheat (Triticum aestivum L.) as a test crop, but only three sites were selected for model calibration and evaluation. The period of simulation for the whole wheat growing season was from 1 April 2004 to 30 July 2004 and detailed analyses were performed for all sites. SWAP simulated soil water dynamics well and the distributed SWAP model is a useful tool to analyze all water balance components.     

Scheduling water application on drip irrigated sugarcane

Irrigation is necessary in order to produce sugarcane in semiarid south Texas, but water supplies are becoming increasing limited. Drip and sprinkler irrigation systems offer more precise water control than conventional furrow irrigation, but are more expensive. This study was conducted to evaluate four different methods (pan evaporation (E_pan), evapotranspiration (ET), auto-tensiometers, manual tensiometers) for determining the amount of irrigation water to apply, and three different frequencies of water application on sugarcane, in order to make the most efficient use of available water using subsurface drip irrigation. The study was conducted over three sugarcane seasons: the plant crop and two ratoon crops. The amount of water applied based on the different methods varied from year to year, with the ET method prescribing the most water in the first ratoon crop but the least amount in the second ratoon. This was probably caused mostly by differences in annual weather conditions. The more frequently water was applied, the larger amount any method tended to prescribe, since more frequent applications resulted in keeping the soil profile fuller, therefore providing less capacity to store rainfall when it occurred. Number of stress days as determined by calculating a stress coefficient based on ET and soil water balance indicated a large amount of stress in the first ratoon but almost none in the second ratoon crop. Direct soil water monitoring indicated much less stress than the calculated levels. Growth measurements and sugarcane yields showed that the highest water applications resulted in the best responses, regardless of the scheduling method used. All irrigation scheduling methods were effective, prescribing similar amounts of water for a given season. Direct measurement using tensiometers gives the most accurate assessment of field conditions, but is expensive and labor intensive. Automated tensiometers were not very reliable. Pan evaporation and ET are effective once they are properly calibrated by developing appropriate coefficients for a particular region. Pan evaporation has been used for a long time, but it is more difficult to obtain reliable data compared to ET data from automated weather stations.     

Comparing Penman–Monteith and Priestley–Taylor approaches as reference-evapotranspiration inputs for modeling maize water-use under Mediterranean conditions

A comparison between experimental and simulated data, considering the Priestley–Taylor (PT) and Penman–Monteith (PM) reference-evapotranspiration (ET₀) approaches was carried out. Experimental data, obtained from an irrigation assessment, conducted during the 1995 and 1996 maize growth-seasons at Zaragoza, Spain, was compared to the mechanistic-model SWAP simulation-results, considering each of the ET₀ calculation approaches in the model input. Soil hydraulic properties, meteorological data, seeding and harvest dates, crop water management and other experimental data were used as SWAP input. As corresponding to the windy and dry conditions found in many Mediterranean landplanes, PT-ET₀ values were significantly lower than PM-ET₀ calculations. Furthermore, simulated actual evapotranspirations considering the PT approach (PT-ET_c) were lower than those found in the simulations that consider the PM approach (PM-ET_c). Correspondingly, simulated drainage flux and soil water contents were higher when the PT-ET₀ approach was used. The correlation coefficients between simulated and measured actual maize evapotranspirations and soil water contents were statistically significant, but the same for both ET₀ calculation approaches. Mean and median differences between actual and simulated maize water-use were not statistically different from zero for both considered ET₀ calculation approaches. Experimental data variability was significantly higher than simulated variability. The comparisons among the evaluated irrigation options, made with the experimental water-use data, lead almost to the same conclusions than those achieved from the simulated maize water-use. Considering PM-ET_c rather than PT-ET_c yields no statistical difference in the modeling-based conclusions. According to the obtained results, the PT approach could be used under Mediterranean conditions for comparative assessments aimed to support irrigation decision-making.     

Reuse of drainage water in irrigation with the aid of 0-1 linear programming

Water management for irrigation in areas with high water scarcity includes not only domestic wastewater treatment but also practices for the reuse of drainage water during the irrigation period. The main problem that concerns the reuse of drainage water for irrigation is the accumulation of salts due to the effluents existing in the soil. In this paper an optimization technique is proposed for the management of drainage water that uses, in combination, a soil-water-plant model (SWAP) and a mixed 0-1 linear programming method. The optimization routine was applied to the irrigation network of Alfeios River in Western Greece, an area that is characterized by high precipitation imbalances between winter and summer months.