The standardized precipitation evapotranspiration index ( SPEI ) was developed in 2010 and has been used in an increasing number of climatology and hydrology studies. The objective of this article is to describe computing options that provide flexible and robust use of the SPEI . In particular, we present methods for estimating the parameters of the log‐logistic distribution for obtaining standardized values, methods for computing reference evapotranspiration ( ET 0 ), and weighting kernels used for calculation of the SPEI at different time scales. We discuss the use of alternative ET 0 and actual evapotranspiration ( ET a ) methods and different options on the resulting SPEI series by use of observational and global gridded data. The results indicate that the equation used to calculate ET 0 can have a significant effect on the SPEI in some regions of the world. Although the original formulation of the SPEI was based on plotting‐positions Probability Weighted Moment ( PWM ), we now recommend use of unbiased PWM for model fitting. Finally, we present new software tools for computation and analysis of SPEI series, an updated global gridded database, and a real‐time drought‐monitoring system.
Actual evapotranspiration (ET ) was measured at 30-min resolution over a 19-month period (September 28, 2000–April 23, 2002) from a nonirrigated pasture site in Florida, USA, using eddy correlation methods. The relative magnitude of measured ET (about 66% of long-term annual precipitation at the study site) indicates the importance of accurate ET estimates for water resources planning. The time and cost associated with direct measurements of ET and the rarity of historical measurements of ET make the use of methods relying on more easily obtainable data desirable. Several such methods (Penman–Monteith (PM), modified Priestley–Taylor (PT), reference evapotranspiration (ET ), and pan evaporation ( )) were related to measured ET using regression methods to estimate PM bulk surface conductance, PT , ET vegetation coefficient, and pan coefficient. The PT method, where the PT is a function of green-leaf area index (LAI) and solar radiation, provided the best relation with ET (standard error (SE) for daily ET of 0.11 mm). The PM method, in which the bulk surface conductance was a function of net radiation and vapor-pressure deficit, was slightly less effective (SE=0.15 mm) than the PT method. Vegetation coefficients for the ET method (SE=0.29 mm) were found to be a simple function of LAI. Pan coefficients for the method (SE=0.40 mm) were found to be a function of LAI and . Historical or future meteorological, LAI, and pan evaporation data from the study site could be used, along with the relations developed within this study, to provide estimates of ET in the absence of direct measurements of ET . Additionally, relations among PM, PT, and ET methods and ET can provide estimates of ET in other, environmentally similar, pasture settings for which meteorological and LAI data can be obtained or estimated.
By various measures (drought area(1) and intensity(2), climatic aridity index(3), and climatic water deficits(4)), some observational analyses have suggested that much of the Earth's land has been drying during recent decades, but such drying seems inconsistent with observations of dryland greening and decreasing pan evaporation(5). 'Offline' analyses of climate-model outputs from anthropogenic climate change (ACC) experiments portend continuation of putative drying through the twenty-first century(3,6-10), despite an expected increase in global land precipitation(9). A ubiquitous increase in estimates of potential evapotranspiration (PET), driven by atmospheric warming(11), underlies the drying trends(4,8,9,12), but may be a methodological artefact(5). Here we show that the PET estimator commonly used (the Penman-Monteith PET13 for either an open-water surface(1,2,6,7,12) or a reference crop(3,4,8,9,11)) severely overpredicts the changes in non-water-stressed evapotranspiration computed in the climate models themselves in ACC experiments. This overprediction is partially due to neglect of stomatal conductance reductions commonly induced by increasing atmospheric CO2 concentrations in climate models(5). Our findings imply that historical and future tendencies towards continental drying, as characterized by offline-computed runoff, as well as other PET-dependent metrics, may be considerably weaker and less extensive than previously thought.
Evapotranspiration has a significant role in agricultural and forest meteorology research, the hydrological cycle, irrigation scheduling and water resources management. Several models are available to estimate evapotranspiration, including mass transfer‐based, radiation‐based, temperature‐based and pan evaporation‐based models. This study aims to assess temperature‐based models versus the Food and Agriculture Organization of the United Nations (FAO) Penman–Monteith model to detect the best one using linear regression under different weather conditions. For this purpose, weather data were gathered from 181 synoptic stations in 31 provinces of Iran. Evapotranspiration was estimated using 11 temperature‐based models and was compared with the FAO Penman–Monteith model. The results showed that the modified Hargreaves–Samani 1 estimates the evapotranspiration better than other models in most provinces of Iran. However, the R 2 values were <0.9930 for 20 provinces of Iran. The best precise method was the modified Hargreaves–Samani 4 for Alborz province (AL). Finally, a list of the best performances of each model was presented to use in other regions according to mean, maximum and minimum temperature elevation, minimum and mean relative humidity, sunshine, precipitation and wind speed. The best weather conditions for use in temperature‐based equations (based on the performance of all methods) are 12–18 °C, 18.0–22.5 °C, 5–13 °C, 40–55%, 2.00–3.25 m s −1 and 230–260 h month −1 for mean, maximum and minimum temperatures, relative humidity, wind speed and sunshine respectively. Results are also useful for selecting the best model when researchers must apply temperature‐based models on the basis of available data.
Mapping evapotranspiration at high resolution with internalized calibration (METRIC) is a satellite-based image-processing model for calculating evapotranspiration (ET) as a residual of the surface energy balance. METRIC uses as its foundation the pioneering SEBAL energy balance process developed in The Netherlands by Bastiaanssen, where the near-surface temperature gradients are an indexed function of radiometric surface temperature, thereby eliminating the need for absolutely accurate surface temperature and the need for air-temperature measurements. The surface energy balance is internally calibrated using ground-based reference ET to reduce computational biases inherent to remote sensing-based energy balance and to provide congruency with traditional methods for ET. Slope and aspect functions and temperature lapsing are used in applications in mountainous terrain. METRIC algorithms are designed for relatively routine application by trained engineers and other technical professionals who possess a familiarity with energy balance and basic radiation physics. The primary inputs for the model are short-wave and long-wave (thermal) images from a satellite (e.g., Landsat and MODIS), a digital elevation model and ground-based weather data measured within or near the area of interest. ET “maps” (i.e., images) via METRIC provide the means to quantify ET on a field-by-field basis in terms of both the rate and spatial distribution. METRIC has some significant advantages over many traditional applications of satellite-based energy balance in that its calibration is made using reference ET, rather than the evaporative fraction. The use of reference ET for the extrapolation of instantaneous ET from periods of 24 h and longer compensates for regional advection effects by not tying the evaporative fraction to net radiation, since ET can exceed daily net radiation in many arid or semi-arid locations. METRIC has some significant advantages over conventional methods of estimating ET from crop coefficient curves in that neither the crop development stages, nor the specific crop type need to be known with METRIC. In addition, energy balance can detect reduced ET caused by water shortage.
Accurate quantification of terrestrial evapotranspiration (ET) is essential to understand the Earth's energy and water budgets under climate change. However, despite water and carbon cycle coupling, there are few diagnostic global evapotranspiration models that have complete carbon constraint on water flux run at a high spatial resolution. Here we estimate 8-day global ET and gross primary production (GPP) at 500 m resolution from July 2002 to December 2017 using a coupled diagnostic biophysical model (called PML-V2) that, built using Google Earth Engine, takes MODIS data (leaf area index, albedo, and emissivity) together with GLDAS meteorological forcing data as model inputs. PML-V2 is well calibrated against 8-day measurements at 95 widely-distributed flux towers for 10 plant functional types, indicated by Root Mean Square Error (RMSE) and Bias being 0.69 mm d and −1.8% for ET respectively, and being 1.99 g C m d and 4.2% for GPP. Compared to that performance, the cross-validation results are slightly degraded, with RMSE and Bias being 0.73 mm d and −3% for ET, and 2.13 g C m d and 3.3% for GPP, which indicates robust model performance. The PML-V2 products are noticeably better than most GPP and ET products that have a similar spatial resolution, and suitable for assessing the influence of carbon-induced impacts on ET. Our estimates show that global ET and GPP both significantly ( < 0.05) increased over the past 15 years. Our results demonstrate it is very promising to use the coupled PML-V2 model to improve estimates of GPP, ET and water use efficiency, and its uncertainty can be further reduced by improving model inputs, model structure and parameterisation schemes.
AbstractQuantification of crop response to the amount of water applied, available, or used is important for decision making to ensure effective, profitable, conservative agricultural production. However, these variables and responses may have interannual attributes and long-term research has rarely quantified interannual variations of irrigation-yield production functions (IYPF), evapotranspiration-yield production functions (ETYPF), and yield response factors (Ky). This long-term research measured grain yield, actual crop evapotranspiration (ETa), increase in ETa attributable to various irrigation levels, basal evapotranspiration (ETb; ET required to establish grain yield), IYPF, ETYPF, and seasonal Ky for maize (Zea mays L.) from 2005 to 2010 growing seasons. Four full and limited irrigation levels [fully irrigated (FIT), 75% FIT, 60% FIT, and 50% FIT] and rainfed treatment were imposed. Seasonal ETa increased linearly with increasing irrigation and the slopes of the ETa versus seasonal irrigation relationships exhibited substantial interannual variation. All-treatment average ETa values were 561, 583, 592, 660, 591, and 628 mm from 2005 to 2010, respectively. Irrigation amounts significantly impacted grain yield in all years with considerable variation among seasons. All-treatment average yield increases attributable to irrigation were 7.7, 4.6, 1.6, 5.8, and 2.7 ton/ha, relative to rainfed treatment, from 2005 through 2010, respectively; six-year average yield increase was 4.5 t/ha. Grain yield had a curvilinear relationship with seasonal irrigation amounts (R2=0.79) and yield increased with irrigation up to approximately 180 mm of irrigation water (15.5 t/ha grain yield); thereafter, irrigation became excessive and diminishing returns occurred. Interannual variation of grain yield produced per unit of irrigation was observed owing to differences in rainfed yield response to precipitation. Totals of 0.92, 1.72, 0.09, 1.06, 1.90, and 0.08 t/ha grain yield were produced per 25.4 mm of irrigation applied (beyond the intercept) from 2005 through 2010, respectively. Based on the pooled (average of all years) curvilinear IYPF equation, 25.4 mm of irrigation produced approximately 1.45 t/ha grain yield [beyond intercept (9.29 t/ha)]. Grain yield had a very strong, linear increase with seasonal ETa (R2>0.92). The ETYPFs exhibited less interannual variation than the IYPFs and the slope of the ETYPFs ranged from 0.0336 in 2008 to 0.0662 in 2010, with the wettest year (2008) having the lowest slope. Six-year average ETYPF had a high R2 (0.93) with an average slope of 0.0409. ETb of individual years also exhibited substantial interannual variation: 263, 319, 314, 209, 319, and 418 mm for the 2005, 2006, 2007, 2008, 2009, and 2010 growing seasons, respectively. Six-year average pooled data indicated that 279 mm of ETb is needed to establish grain yield. Based on the ETYPF data from individual years, 25.4 mm of ETa resulted in 0.96, 1.20, 1.14, 0.85, 1.22, and 1.68 t/ha maize yield (beyond the intercept) from 2005 through 2010. When individual treatments were considered, 25.4 mm of ETa resulted in 0.98, 0.40, 0.62, 1.01, and 0.48 t/ha grain yield (beyond intercept) for rainfed, 50% FIT, 60% FIT, 75% FIT, and FIT, respectively. The climatic conditions in August were the most influential in determining the slope of the ETYPFs. Seasonal average Ky also exhibited interannual variation, which was 1.89, 1.89, 1.39, 1.80, and 2.64 in 2006, 2007, 2008, 2009, and 2010, respectively. This research provides important data, information, and analyses in terms of interannual variation of various parameters on maize response to water.
Partitioning of evapotranspiration ( ) into evaporation from the soil ( ) and transpiration through the stomata of plants ( ) is challenging but important in order to assess biomass production and the allocation of increasingly scarce water resources. Generally, is the desired component with the water being used to enhance plant productivity; whereas, is considered a source of water loss or inefficiency. The magnitude of is expected to be quite significant in sparsely vegetated systems, particularly in dry areas or in very wet systems such as surface irrigated crops and wetlands. In these cases, partitioning is fundamental to accurately monitor system hydrology and to improve water management practices. This paper aims to evaluate and summarize available methods currently used to separately determine and components. We presuppose that, to test the accuracy of partitioning methods (measurements and/or modeling), all three components, i.e., , and , must be estimated independently, but recognize that sometimes one of the components is taken as the residual of the other two. Models that were validated against measurements for their ability to partition between and are briefly discussed. To compare approaches, 52 partitioning studies were considered regarding estimates of the relative amount of and for success of agreement in closing the = + equation. The / ratio was found to exceed 30% in 32 of the studies, which confirms the hypothesis that often constitutes a large fraction of and deserves independent consideration. Only 20 studies estimated and as well as , and had varied results. A number of studies succeeded to estimate + to within 10% of measured . Future challenges include development of models simulating the components of separately and advancement of methods for continuous measurement of , and/or the ratio between the two.