Artificial neural network estimation of reference evapotranspiration from pan evaporation in a semi-arid environment

The objective of this study was to test an artificial neural network (ANN) for converting pan evaporation data (E _p) to estimate reference evapotranspiration (ET₀) as a function of the maximum and minimum air temperature. The conventional method that uses Pan coefficient (K _p) as a factor to convert E _p to ET₀, is also considered for the comparison. The ANN has been evaluated under semi-arid conditions in Safiabad Agricultural Research Center (SARC) in the southwest of Iran, comparing daily estimates against those from the FAO-56 Penman–Monteith equation (PM), which was used as standard. The comparison shows that, the conventional method underestimated ET₀ obtained by the PM method. The ANN method gave better estimates than the conventional method that requires wind speed and humidity data.     

Suspended sediment load prediction of river systems: An artificial neural network approach

Information on suspended sediment load is crucial to water management and environmental protection. Suspended sediment loads for three major rivers (Mississippi, Missouri and Rio Grande) in USA are estimated using artificial neural network (ANN) modeling approach. A multilayer perceptron (MLP) ANN with an error back propagation algorithm, using historical daily and weekly hydroclimatological data (precipitation P_(t), current discharge Q_(t), antecedent discharge Q_(t−1), and antecedent sediment load SL_(t−1)), is used to predict the suspended sediment load SL_(t) at the selected monitoring stations. Performance of ANN was evaluated using different combinations of input data sets, length of record for training, and temporal resolution (daily and weekly data). Results from ANN model were compared with results from multiple linear regressions (MLR), multiple non-linear regression (MNLR) and Autoregressive integrated moving average (ARIMA) using correlation coefficient (R), mean absolute percent error (MAPE) and model efficiency (E). Comparison of training period length was also made (4, 3 and 2 years of training and 1, 2 and 3 years of testing, respectively). The model efficiency (E) and R² values were slightly higher for the 4 years of training and 1 year of testing (4 * 1) for Mississippi River, indifferent for Missouri and slightly lower for Rio Grande River. Daily simulations using Input 1 (P_(t), Q_(t), Q_(t−1), SL_(t−1)) and three years of training and two years of testing (3 * 2) performed better (R² and E of 0.85 and 0.72, respectively) than the simulation with two years of training and three years of testing (2 * 3) (R² and E of 0.64 and 0.46, respectively). ANN predicted daily values using Input 1 and 3 * 2 architecture for Missouri (R² = 0.97) and Mississippi (R² = 0.96) were better than those of Rio Grande (R² = 0.65). Daily predictions were better compared to weekly predictions for all three rivers due to higher correlation within daily than weekly data. ANN predictions for most simulations were superior compared to predictions using MLR, MNLR and ARIMA. The modeling approach presented in this paper can be potentially used to reduce the frequency of costly operations for sediment measurement where hydrological data is readily available.     

Partitioning of seasonal evapotranspiration from a commercial furrow-irrigated Sultana vineyard

Seasonal partitioning of evapotranspiration (ET) between transpiration by grapevines (Vitis vinifera) (T _gp) and by cover crops of a ryegrass/clover mixture (T _cc), and soil evaporation (E _s) was performed for a furrow-irrigated vineyard during the 1994/1995 and 1995/1996 growing seasons in south-eastern Australia. ET, determined with a water balance approach, averaged 622 mm. The ET rate averaged over the two seasons increased from 2 mm day^–1 in spring (September to November), when it was dominated by T _cc, to peak rates of around 5 mm d^–1 in summer (December to February) when it was dominated by E _s. T _gp, determined with either heat-pulse sensors or the Penman-Monteith equation, attained peak rates of 0.75 and 0.98 mm d^–1, or 6.2 and 8.1 l vine^–1 day^–1 in the first and second seasons, respectively. Total seasonal T _gp of 109.1 mm (900 l vine^–1) in 1994/1995 and 118.8 mm (980 l vine^–1) in 1995/1996 constituted just 18 – 19% of total ET. T _cc totalled 214 mm (34% of ET) in the first season, when pasture cover was sparse and present for 5 months of the growing season (September to February), and 196 mm (30% of ET) in the second season when pasture cover was heavy but present for only 3 months (September to November). E _s averaged 49% of total ET over both seasons. At least 30% of water used for ET was contributed by antecedent soil water in both seasons. The crop factor (K _c) was largely constant throughout the season with an average value of 0.48. The depletion pattern of soil water indicated that the vine explored the soil profile well beyond 1.0 mm depth and almost evenly up to a distance of 1.5 m from the trunk. Water use efficiencies for fresh fruit yield (WUE), i. e., the ratio of fruit weight to total water use at harvest,were 13.3 and 40.5 kg ha^–1 mm^–1 when based on ET in 1994/1995 and 1995/1996, respectively, and 84.0 and 211.1 kg ha^–1 mm^–1, respectively, when based on T _gp. The T _gp data were used to verify three models of vine transpiration developed in an earlier study. Models based on the green area index or on fraction of incident radiation intercepted by the vine canopy produced good agreement with T _gp. The model based on canopy resistance performed poorly, indicating the difficulty of extrapolating the stomatal response to environmental variables from one set of experimental conditions to another. Received: 23 September 1996     

Sprinkler irrigation scheduling of winter wheat in the North China Plain using a 20 cm standard pan

Based on evaporation from a 20 cm diameter pan placed above the crop canopy, sprinkler irrigation scheduling of winter wheat was studied in the North China Plain (NCP) in the 2001–2004 winter wheat seasons. Results showed that pan evaporation (E _pan,C) was closely related to actual evapotranspiration (ET) measured using weighing lysimeters. The combined pan–crop coefficient (K _c,pan), the ratio of ET to E _pan,C, was closely related to leaf area index (LAI ) and plant height. Data from the 2002–2003 season were used to establish the relationships between K _c,pan and LAI (method A) or plant height (method B), and used to determine the crop coefficient (method C). ET computed by the three methods was compared with measured ET using lysimeters in the 2001–2002 and 2003–2004 seasons. Mean relative error of estimated daily ET by the three methods ranged from 20 to 30%, and the relative error in cumulative ET in the experimental periods ranged from 1 to 19%. Among the three methods, results from methods A and B were not significantly different from each other (P > 0.01), and were closer to the lysimeter data than results from method C (P < 0.001). Method B, being easier to measure, was recommended for ET estimation in NCP.     

Relating leaf area index of natural eucalypt vegetation to climate variables in southern Australia

We address a need for a rapid technique to estimate the leaf area index (LAI) of pre-existing natural vegetation. This is required to determine the effects of agroforestry plantings on deep drainage from agricultural land. Previous work shows: (1) a relationship between the hydrologic ‘footprint’ of tree belts and their lineal leaf area (leaf area per metre of belt, LLA, m² m⁻¹), relative to the LAI of natural vegetation and (2) that the LAI of natural vegetation is related to plant-available soil moisture and climate. We evaluated relationships between LAI measured at 37 sites across southern Australia and (1) annual average rainfall, P; (2) annual average pan evaporation, E₀; (3) ‘available rainfall’—annual average rainfall, P, minus annual average pan evaporation, E₀; (4) a climate wetness index P/E₀; (5) Specht's soil evaporative index, k = 0.0045 + 71.57/E₀. P was the best indicator for the data set used, i.e. LAI = 0.003P + 0.41 (r² = 0.80).     

Measurement and simulation of evaporation from soil in olive orchards

Evaporation from the soil (E _s) beneath an olive orchard was characterised in a semi-arid Mediterranean climate (Córdoba, Spain). First, the microlysimeter method was modified to measure accurately E _s beneath tree orchards. The variability in irradiance reaching the soil beneath the orchard caused spatial variations in E _s during both evaporation stages. In the first days of the drying cycle, E _s was higher for high irradiance locations but the opposite occurred the subsequent days, although daily differences in E _s between locations progressively declined. For the energy-limiting stage, linear relationships between E _s values and incident photosynthetically active radiation were found for different times throughout the season. The slopes of the relationships were similar, but their intercepts differed substantially, showing the importance of a variable aerodynamic component in determining E _s. A simple functional model was formulated to estimate E _s at daily time steps. During the energy-limiting stage, E _s is calculated as the sum of the equilibrium evaporation at the soil surface and an aerodynamic term, derived from the Penman equation. For the falling rate stage, Ritchie's (1972) approach is adopted for the E _s calculations. The model was successfully tested in an orchard of 6×6 m spacing, typical of intensive olive orchards, under a wide range of evaporative demand conditions. Trees covered around 36% of the soil surface. The model predicted an average seasonal E _s of 286 mm, which represents around one third of the estimated olive evapotranspiration and about 50% of the average seasonal rainfall of the area. Received: 3 August 1998     

Characterisation of water use by Sultana grapevines (Vitis vinifera L.) on their own roots or on Ramsey rootstock drip-irrigated with water of different salinities

Seasonal evapotranspiration (ET) was determined for Sultana grapevines grown on their own roots (Own-rooted) or grafted onto Ramsey rootstock (Grafted), and irrigated with water of three salinity levels – low (0.4 dS m^–1), medium (1.8 dS m^–1) and high (3.6 dS m^–1) – during the 1994/1995 growing season in south-eastern Australia. Transpiration (T) was determined from sap flux, soil evaporation (E _s) with a model, and soil water (S) with a neutron probe. Total ET for the season was similar for both Own-rooted and Grafted, averaging 382 mm. However, Grafted partitioned a mean of 193.5 mm (50.8%) of the ET through T compared to 146.7 mm (38.4%) by Own-rooted. Daily rates of T were generally low and attained peaks of 1.2 mm (9.9 l per vine) for Grafted and 0.9 mm (7.5 l) for Own-rooted in late November, and changed very little until after harvest in February. In contrast to T, the E _s rate was consistently higher for Own-rooted than for Grafted from November onwards, and at the end of the season totalled 237 mm for Own-rooted compared to 187 mm for Grafted. Differences between Own-rooted and Grafted in their partitioning of ET into T and E _s were associated with their canopy development. Grafted had a higher rate of canopy development than Own-rooted, and in mid-season, the former intercepted about 50% more incident radiation than Own-rooted. The crop factors, i. e. ratio of water use to evaporative demand, based on ET were similar for both vine types with an average seasonal value of 0.25, but when based on T were 0.12 for Grafted and 0.10 for Own-rooted. The ratio of fresh fruit weight to total water used at harvest, i. e. crop water use efficiency (CWUE), based on ET, had a mean of 86 kg mm^–1 ha^–1 for Grafted and 43 kg mm^–1 ha^–1 for Own-rooted, and when based on T, was 165 and 115 kg mm^–1 ha^–1, respectively; however, supplementary data obtained during the 1993/1994 season, indicated a CWUE based on T of 294 and 266 kg mm^–1 ha^–1 for Grafted and Own-rooted, respectively. Salinity did not have significant effects on canopy development and water use for most of the 1994/1995 growing season. The study shows ET and crop factors for the drip-irrigated grapevines to be much lower than previously reported for this district. Received: 6 May 1996     

Maximum daily trunk shrinkage and stem water potential reference equations for irrigation scheduling of lemon trees

Measurements of midday stem water potential (ψ_stem) and maximum daily trunk shrinkage (MDS) were done over a 3-year period in adult Fino lemon trees (Citrus limon (L.) Burm. fil.) grafted on sour orange (C. aurantium L.) rootstocks. Plants were irrigated daily above their water requirements in order to obtain non-limiting soil water conditions. The results indicated that reference equations can be obtained for MDS and ψ_stem by pooling data across several seasons using crop reference evapotranspiration (ETo), daily mean vapor pressure deficit (VPD_m) and mean daily air temperature (T _m) in the case of MDS, and ETo in the case of ψ_stem. The best predictor of MDS under non-limiting soil water conditions was T _m, suggesting that MDS reference values can be obtained by means of easy and cheap measurements. MDS and ψ_stem values were not influenced significantly by yield or crop load variations between years. A negative linear relationship between MDS and ψ_stem was found, pointing to an unchanging radial hydraulic conductivity in the bark tissues and suggesting that the MDS is controlled by water potential.     

Simulation of artificial neural network model for trunk sap flow of Pyrus pyrifolia and its comparison with multiple-linear regression

Trunk sap flow of tree is an important index in the irrigation decision of orchard. On the basis of the measured sap flow (SF) of pear tree (Pyrus pyrifolia) in the field, the multiple-linear regression for simulating the SF was obtained after analyzing the relationships between the SF and its affecting factors in this study and an artificial neural network (ANN) technique was applied to construct a nonlinear mapping to simulate the SF, then the simulated SF by two models was, respectively, compared to the measured value. Results showed that trunk SF had significant relationship with the vapour pressure deficit (VPD) in the single-variable analysis method but with soil volumetric water content (θ) using the ANN models with default of different variables. The correlation coefficient (R²), mean relative error (MRE) and root mean square error (RMSE) between the measured and simulated sap flows by the ANN model developed by taking VPD, solar radiation (S_r), air temperature (T), wind speed (W_s), θ, leaf area index (LAI) as the input variables were 0.953, 10.0% and 5.33 L d⁻¹, respectively, and the simulation precision of ANN model was superior to that of multiple-linear regression due to its better performance for the nonlinear relationship between trunk SF and its affecting factors, thus ANN model can simulate trunk sap flow and then may help the efficient water management of orchard.     

Alternative models for predicting the foliage—Air temperature difference of well irrigated wheat under variable meteorological conditions. I. Derivation of parameters II. Accuracy of predictions

Summary One means of using infrared measurements of foliage temperature (T _f ) for scheduling irrigations requires the use of meteorological data to predict the foliage-air temperature difference for a comparable well-watered crop (T _f ^* – T _a ). To determine the best method for making this prediction, parameters for models of increasing complexity for predicting (T _f ^* – T _a ) were derived for wheat using two sets of field data collected in 1982 and 1983.The simplest model with vapor pressure deficit (VPD) as the sole predictor accounted for 64% of observed variance in (T _f ^* – T _a ). The next model with both VPD and net radiation (R _n ) as predictors accounted for 74%. The most complex model predicted (T _f ^* – T _a ) from the crop energy balance. In addition to VPD and R _n it included parameters for the effects of air temperature (T _a ), aerodynamic resistance (r _a ) and the canopy resistance of a well-watered crop (r _cp ) and accounted for 70% of the variance.Accuracy of these alternative models was tested against an independent set of field data collected in 1984. The single variable model with VPD as sole predictor accounted for 17% of the variance in observed values of (T _f ^* – T _a ). This increased to 47% when the effect of R _n was included by using the two variable model and was increased further to 65% when the additional variables of T _a , r _a and r _cp were included by use of the energy balance model. When the complexity of the model was measured by its number of variables there was a close relationship between complexity and the accuracy of the predictions. Reasons for the residual variability are discussed. The need for improved instrumentation for meteorological measurements was indicated.     

Evapotranspiration predictions: a comparison among GLEAMS,Opus, PRZM-2, and RZWQM models in a humid and thermic climate

Environmental fate models are increasingly used to evaluate potential impacts of agrochemicals on water quality to aid in decision making. However, errors in predicting processes like evapotranspiration (ET), which is rarely measured during model validation studies, can significantly affect predictions of chemical fate and transport. This study compared approaches and predictions for ET by GLEAMS, Opus, PRZM-2, and RZWQM and determined effects of the predicted ET on simulations of other hydrology components. The ET was investigated for 2 years of various fallow–corn growing seasons under sprinkler irrigation. The comparison included annual cumulative daily potential ET (ET_p), actual ET, and partitioning of total ET between soil evaporation (E_s) and crop transpiration (E_t). When measured pan evaporation was used for calculating ET_p (the pan evaporation method), Opus, PRZM-2, and RZWQM predicted 74, 65, and 59%, respectively, of the 10-year average ET reported for a nearby site. When the energy-balance equations were used for calculating ET_p (the combination methods), GLEAMS, Opus, PRZM-2, and RZWQM predicted 84, 105, 60, and 72% of the reported ET, respectively. The pan evaporation method predicted a similar amount of ET to the combination methods for bare soil, but predicted less ET when both E_s and E_t occurred. RZWQM reasonably predicted partitioning of ET to E_s, while GLEAMS and Opus over-predicted this partitioning. A close correlation between soil water storage in the root zone and ET suggests that accurate soil water content predictions were fundamental to ET predictions. ©     

Components of the water balance in soil with sugarcane crops

The objective of this study was to analyze the components of the water balance in an Ultisol, located in the municipality of Jaboticabal, SP, Brazil (21°20′20″S, 48°18′35″W), that was cultivated with sugarcane. The monitoring was performed during the agricultural cycle of the first ratoon between 11/16/2006 and 7/9/2007. Three treatments were established in four blocks with doses of ammonium sulfate, as follows: Treatment 1 (T₁), without fertilizer; Treatment 2 (T₂), 100 kg ha⁻¹ of nitrogen (N) and 114 kg ha⁻¹ of sulfur (S); and Treatment 3 (T₃), 150 kg ha⁻¹ of N and 172 kg ha⁻¹ of S. Rainy precipitation (P) in the area was measured with a rain gauge. The soil water storage (H) and the soil water storage variations (ΔH) were determined by the gravimetric method, and the internal drainage (D)/capillary rise (CR) at a depth of 0.9 m was quantified by the water flux density using the Darcy–Buckingham equation. The actual evapotranspiration (ET_a) was calculated as follows: ET_a = P − D + CR ± ΔH. During the study period, the amount of rainfall was 1406 mm, 121 mm greater than the historic average for the region (1285 mm), with a notable peak in the month of January of 402 mm (historic average: 251 mm). The internal drainage was 300 mm under T₁, 352 mm under T₂, and 199 mm under T₃, and this was relevant during times with elevated P, when the actual H was greater than the field capacity H. The actual evapotranspiration (T₁: −897.7 mm, T₂: −847.5 mm, and T₃: −970.8 mm) and the water use efficiency (T₁: −131.3 kg mm⁻¹, T₂: −146.6 kg mm⁻¹, and T₃: −127.5 kg mm⁻¹) did not differ among the treatments. The dispersion of D was greater than the other components of the water balance, especially during the period of elevated P, with the errors of this process propagated in the estimation of ET_a. Despite of this propagated standard deviation of ET_a, it accounted less than 15% of the total ET_a, showing that the method may be conveniently used in field studies with sugarcane crops.     

Least squares support vector machine for modeling daily reference evapotranspiration

The accuracy of a least square support vector machine (LSSVM) in modeling of reference evapotranspiration (ET₀) was examined in this study. The daily weather data, solar radiation, air temperature, relative humidity and wind speed of two stations, Glendale and Oxnard, in southern district of California, were used as inputs to the LSSVM models to estimate ET₀ obtained using the FAO-56 Penman–Monteith equation. In the first part of the study, LSSVM estimates were compared with those of the following empirical models: Priestley–Taylor, Hargreaves and Ritchie methods. The comparison results indicated that the LSSVM performed better than the empirical models. In the second part of the study, the LSSVM results were compared with those of the conventional feed-forward artificial neural networks (ANN). It was found that the LSSVM models were superior to the ANN in modeling ET₀ process.     

General irrigation efficiency for field water management

On-farm water management systems are traditionally evaluated using a set of performance indices which are inconvenient for evaluation and comparison. We propose a general efficiency (E_g) which is defined as the ratio of crop transpiration to the sum of the volume of applied water and the volume of deficit. E_g combines the characteristics of traditionally used irrigation efficiencies: application efficiency (E_a), storage efficiency (E_s) and the Christiansen's coefficient of uniformity (U_c). Thus, it is possible to compare the performance of different water application systems and/or design and management scenarios using a single index. The relationships of E_g vs. E_a, E_s and U_c are also presented using a transpiration fraction (α) which is defined as the ratio of transpiration to evapotranspiration.     

Water requirements of olive orchards: I simulation of daily evapotranspiration for scenario analysis

Water requirements of olive orchards are difficult to calculate, since they are influenced by heterogeneous factors such as age, planting density and irrigation systems. Here we propose a model of olive water requirements, capable of separately calculating transpiration (E _p), intercepted rainfall evaporation (E _pd) and soil evaporation (E _s) from the wet and dry fraction of the soil surface under localized irrigation. The model accounts for the effects of canopy dimension on E _p and of the wetted soil surface fraction on E _s. The model was tested against actual measurements of olive evapotranspiration (ET) obtained by the eddy covariance technique in a developing olive orchard during 3 years. The predicted ET and crop coefficients showed good agreement with the measured data. The model was then used to simulate the average water requirements of two mature orchards using 20-year meteorological datasets of Cordoba (Spain) and Fresno (CA, USA). Average annual ET of a 300 trees ha⁻¹ orchard at Cordoba was 1,025 mm, while the same orchard at Fresno had an average ET of 927 mm. Transpiration losses were 602 mm at Cordoba and 612 mm at Fresno. Evaporation from the soil can have a large effect on olive ET; thus, olive crop coefficients (K _c) are very sensitive to the rainfall regime.     

Soil evaporation from drip-irrigated olive orchards

Evaporation from the soil (E_s) in the areas wetted by emitters under drip irrigation was characterised in the semi-arid, Mediterranean climate of Córdoba (Spain). A sharp discontinuity in E_s was observed at the boundary of the wet zone, with values decreasing sharply in the surrounding dry area. A single mean value of evaporation from the wet zone (E_sw) was determined using microlysimeters. Evaporation from the wet zones of two drip-irrigated olive orchards was clearly higher than the corresponding values of E_s calculated assuming complete and uniform soil wetting (E_so), demonstrating the occurrence of micro-scale advection in olive orchards under drip irrigation. Measurements over several days showed that the increase in evaporation due to microadvection was roughly constant regardless of location and of the fraction of incident radiation reaching the soil. Thus, daily evaporation from wet drip-irrigated soil areas (E_sw) could be estimated as the sum of E_so and an additive microadvective term (T_MA). To quantify the microadvective effects, we developed variable local advective conditions by locating a single emitter in the centre of a 1.5 ha bare plot which was subjected to drying cycles. E_sw increased relative to E_so as the soil dried and advective heat transfer increased evaporation from the area wetted by the emitter. The microadvective effects on E_s were quantified using a microadvective coefficient (K_sw), defined as the ratio between E_sw and E_so. A model was then developed to calculate T_MA for different environmental and orchard conditions. The model was validated by comparing measured E_sw against simulated evaporation (E_so+T_MA) for different soil positions and environmental conditions in two drip-irrigated olive orchards. The mean absolute error of the prediction was 0.53 mm day^-1, which represents about a 7% error in evaporation. The model was used to evaluate the relative importance of seasonal E_s losses during an irrigation season under Córdoba conditions. Evaporation from the emitter zones (E_sw) represented a fraction of seasonal orchard evapotranspiration (ET), which ranged from 4% to 12% for a mature (36% ground cover) and from 18% to 43% of ET for a young orchard (5% ground cover), depending on the fraction of soil surface wetted by the emitters. Estimated potential water savings by shifting from surface to subsurface drip ranged from 18 to 58 mm in a mature orchard and from 28 to 93 mm in a young orchard, assuming daily drip applications and absence of rainfall during the irrigation season.     

Contrasting methods for estimating evapotranspiration in soybean

Crop scientists are often interested in canopy rather than leaf water estimates. Comparing canopy fluxes for multiple treatments using micrometeorological approaches presents limitations because of the large fetch required. The goal of this study was to compare leaf-scale to field-scale data by summing soil water evaporation (E) and leaf transpiration (T) versus ET using tower eddy covariance (EC) and scaling leaf transpiration to the canopy level using a two-step scaling approach in soybean [Glycine max (L.) Merr.]. Soybean transpiration represented 89-96% of E + T when combining the soil water evaporation with leaf transpiration on the five measurement days during reproductive growth. Comparing E + T versus ET from the EC system, the E + T method overestimated ET from 0.68 to 1.58 mm. In terms of percent difference, the best agreement between the two methods was 15% on DOY 235 and the worst agreement occurred on DOY 234 (41%). A two-step scaling method predicted average ET within 0.01 mm of the EC ET between 10:00 and 14:15 on an hourly time-step on DOY 227 under uniform sky conditions and average ET within 0.03 mm of the EC ET on DOY 235 under intermittent sky conditions between 10:00 and 15:15. Pooling the scaled-leaf data and comparing them with the measured EC ET data exhibited a strong linear relationship (r = 0.835) after accounting for bias (6%). Findings from this study indicate satisfactory results comparing absolute differences are likely not obtainable by summing leaf transpiration with soil water evaporation to calculate canopy water fluxes. However, scaling leaf transpiration provided a robust measure of canopy transpiration during reproductive growth in soybean under these conditions and merits additional study under different climatic and crop conditions.     

Daily pan evaporation modeling using linear genetic programming technique

This paper investigates the ability of linear genetic programming (LGP), which is an extension to genetic programming (GP) technique, in daily pan evaporation modeling. The daily climatic data, air temperature, solar radiation, wind speed, pressure and humidity of three automated weather stations, Fresno, Los Angeles and San Diego in California, are used as inputs to the LGP to estimate pan evaporation. The LGP estimates are compared with those of the Gene-expression programming (GEP), which is another branch of GP, multilayer perceptrons (MLP), radial basis neural networks (RBNN), generalized regression neural networks (GRNN) and Stephens–Stewart (SS) models. The performances of the models are evaluated using root mean square errors (RMSE), mean absolute error (MAE) and determination coefficient (R ²) statistics. Based on the comparisons, it was found that the LGP technique could be employed successfully in modeling evaporation process from the available climatic data.     

Water requirements of olive orchards–II: determination of crop coefficients for irrigation scheduling

Intensification of olive cultivation by shifting a tree crop that was traditionally rain fed to irrigated conditions, calls for improved knowledge of tree water requirements as an input for precise irrigation scheduling. Because olive is an evergreen tree crop grown in areas of substantial rainfall, the estimation of crop evapotranspiration (ET) of orchards that vary widely in canopy cover, should be preferably partitioned into its evaporation and transpiration components. A simple, functional method to estimate olive ET using crop coefficients (K _c=ET/ET₀) based on a minimum of parameters is preferred for practical purposes. We developed functional relationships for calculating the crop coefficient, K _c, for a given month of the year in any type of olive orchard, and thus its water requirements once the reference ET (ET₀) is known. The method calculates the monthly K _c as the sum of four components: tree transpiration (K _p), direct evaporation of the water intercepted by the canopy (K _pd), evaporation from the soil (K _s1) and evaporation from the areas wetted by the emitters (K _s2). The expression used to calculate K _p requires knowledge of tree density and canopy volume. Other parameters needed for the calculation of the K _c’s include the ET₀, the fraction of the soil surface wetted by the emitters and irrigation interval. The functional equations for K _p, K _pd, K _s1 and K _s2 were fitted to mean monthly values obtained by averaging 20-year outputs of the daily time step model of Testi et al. (this issue), that was used to simulate 124 different orchard scenarios.     

Variable upper and lower crop water stress index baselines for corn and soybean

Upper and lower crop water stress index (CWSI) baselines adaptable to different environments and times of day are needed to facilitate irrigation scheduling with infrared thermometers. The objective of this study was to develop dynamic upper and lower CWSI baselines for corn and soybean. Ten-minute averages of canopy temperatures from corn and soybean plots at four levels of soil water depletion were measured at North Platte, Nebraska, during the 2004 growing season. Other variables such as solar radiation (R _s), air temperature (T _a), relative humidity (RH), wind speed (u), and plant canopy height (h) were also measured. Daily soil water depletions from the research plots were estimated using a soil water balance approach with a computer model that used soil, crop, weather, and irrigation data as input. Using this information, empirical equations to estimate the upper and lower CWSI baselines were developed for both crops. The lower baselines for both crops were functions of h, vapor pressure deficit (VPD), R _s, and u. The upper baselines did not depend on VPD, but were a function of R _s and u for soybean, and R _s, h, and u for corn. By taking into account all the variables that significantly affected the baselines, it should be possible to apply them at different locations and times of day. The new baselines developed in this study should facilitate the application of the CWSI method as a practical tool for irrigation scheduling of corn and soybean.