A satellite-based Daily Actual Evapotranspiration estimation algorithm over South Florida

doi:10.1016/j.gloplacha.2008.12.008

Global and Planetary Change

Volume 67, Issues 1–2, May 2009, Pages 62-77

https://doi.org/10.1016/j.gloplacha.2008.12.008 Get rights and content

Abstract

Water resources and agricultural applications require the knowledge of evapotranspiration (ET) over a range of spatial and temporal scales. Due to paucity of surface based hydro-meteorological stations, the spatial resolution of ET estimates is fairly coarse and is not particularly suitable or reliable for hydrologic modeling, water resources planning and decision making. An ET estimation algorithm has been developed by combining data from satellite and ground observations. The method extends the applicability of a commonly used energy balance formulation of ET and utilizes the contextual relationship between remotely sensed surface temperature and vegetation index. The required parameters are derived from the Advanced Very High Resolution Radiometer (AVHRR) aboard NOAA-14 satellite. First, the Evaporative Fraction (EF) is estimated by utilizing the relationship between a vegetation index and radiometric surface temperature observed from AVHRR for each day. Then spatio-temporal interpolation and filtering techniques are applied to obtain daily EF values for cloudy pixels to produce the EF map for the entire region. Daily Actual ET (DAET) maps are derived from these EF maps and net radiation maps obtained from ground-based observations. The comparisons between satellite derived DAET and ground measurements showed overall low bias and root-mean-square-error for both clear and cloudy days at South Florida in 1998 and 1999. The proposed satellite-based DAET (SatDAET) algorithm has its EF component primarily estimated from satellite data and the resulting DAET has been validated using multi-year ground observations over the South Florida region. The SatDAET algorithm appears to be robust and has the potential to provide near real-time land surface evapotranspiration monitoring over large heterogeneous areas at a very fine spatial and temporal resolution.

Introduction

Evapotranspiration (ET) is important for water resources management, hydrometeorological predictions, environmental conservation, and agriculture competitiveness. Agriculture alone in the United States accounts for 80% of the nation's consumptive water use. Crop yield decreases if irrigation and rainfall amounts are not equal to ET, while excessive irrigation may cause percolation of water along with agri-chemicals below vadoze zone. Accurate and temporally continuous ET estimation over large areas will provide valuable assistance to efficient water use and irrigation management.

ET is an indicator of the rate of change of the global water cycle, and is a necessary variable for most numerical weather forecasting and global climate model simulations. Depending on water availability, climate regimes, and landscape conditions, ET can represent a substantial portion of the regional water budget. For example, in comparison to the long-term averaged precipitation of 1346 mm (or 53 in.) per year over the South Florida Everglades, the long-term averaged and spatially variable ET ranges between 889 mm (or 35 in.) and 1397 mm (or 55 in.) thus representing a substantial portion of the water budget for the ecosystem. At present, due to paucity of surface based hydro-meteorological stations all across the globe, the spatial representation of point ET estimates is questionable and is not particularly reliable for spatially distributed hydrometeorological modeling and decision making. Remote sensing provides an economically reliable way to estimate ET over large heterogeneous areas. Over the past decades, use of remotely sensed land surface data was largely limited to applications that directly detect surface or near surface features such as snow/ice mapping, vegetation mapping, and cloud masking.

Modeling approaches for point estimation of ET perform reasonably well for different sites. For distributed ET estimation, however, two major difficulties are: (1) necessity of using surface meteorological observations that are not readily available over large areas, and (2) requirement of a complex land surface model which must be interfaced with remote sensing data. Our proposed methodology is intended to avoid complicated parameterizations that involve heat and momentum resistances to energy transfer for ET estimation, and to avoid numerous correction procedures. This methodology is based on commonly used energy balance formulation of ET (Priestley and Taylor, 1972) and a contextual relationship between remotely sensed temperature and vegetation index. We use the spatial context (i.e., spatial variation of collocated surface processes) from satellite images to estimate ET with minimum number of free parameters. Our recent research efforts have shown that high-resolution (1-km) Daily Actual ET (DAET) mapping of land surface is feasible with Advanced Very High Resolution Radiometer (AVHRR) data from once a day afternoon overpass of polar-orbiting satellite NOAA-14 (Jiang and Islam, 1999, Jiang and Islam, 2001, Islam et al., 2002, Jiang and Islam, 2003).

Our initial algorithm development and application over heterogeneous domain under clear sky conditions has shown the simplicity of the method without compromising the accuracy in results. Application of ET algorithm, for six clear sky days over Southern Great Plains in US in 1997, showed that the overall error in near noon time surface latent heat flux was about 50.5 Wm^− 2 in terms of root mean square error (RMSE) (Jiang and Islam, 2001), which is about 17.6% of the observed mean. Intercomparison of the method to the commonly used aerodynamic resistance energy balance residual method for surface latent heat flux estimation on theoretical basis and real-time performance showed this method requires fewer surface parameters than the traditional approach and can produce comparable or better results (Jiang and Islam, 2003).

The objective of the present study is to explore the applicability of the original Jiang–Islam method to meet the needs of near real-time operational water resources management practice by providing continuous Daily Actual Evapotransipiration (DAET) estimation over the South Florida region for years 1998 and 1999, for both clear and cloudy days. We will introduce a spatial–temporal interpolation scheme that is applicable for efficient utilization of remote sensing signals in potentially near real-time operational mapping of DAET over a heterogeneous domain. The enhanced comprehensive algorithm — SatDAET — will be discussed in detail. Section 2 describes the original method with its enhancements to provide DAET for clear and cloudy days. Section 3 presents the application of SatDAET algorithm for the South Florida Water Management District (SFWMD) domain. Section 4 presents the results of implementing the enhanced algorithm in a continuous fashion over the domain for all days in 1998 and 1999. The considerations on model performance evaluation metrics and validations by available ground-based observation data for both clear and cloudy days of the results are also discussed in this section. Discussion of uncertainty sources and conclusion is presented in Section 5.

Section snippets

Basic framework

ET can be estimated by Priestley–Taylor type models as $E T = ϕ (R_{n} - G) \frac{Δ}{Δ + γ}$ where R_n is the net radiation, G is the soil heat flux, Δ is the slope of saturated vapor pressure at the air temperature, and γ is a psychrometric constant. This equation is applicable for a range of surface conditions where R_n − G is the driving force for ET, $(R_{n} - G) \frac{Δ}{Δ + γ}$ is the equilibrium ET, and ϕ represents a product of Evaporative Fraction (EF) and $\frac{Δ + γ}{Δ}$ . It is important to note that in the absence of significant advection

Remote sensing data processing

The raw remote sensing data were extracted from the High Resolution Picture Transmission (HRPT) AVHRR data available at the previous NOAA/NESDIS' Satellite Active Archive (SAA) — http://www.saa.noaa.gov, which is now migrated into the NOAA's Comprehensive Large Array-data Stewardship System (CLASS) at http://www.osd.noaa.gov/class. Scanline resolution AVHRR afternoon overpass channels 1, 2, and 4 (i.e., VIS, near IR and IR channels respectively) with associated solar and satellite scan angles

Results of DAET maps for clear and cloudy days

Panel Fig. 3, Fig. 5, Fig. 6 show the DAET maps derived using the SatDAET algorithm for clear (Fig. 3), partially cloudy (Fig. 4), cloudy (Fig. 5), and poor image quality (almost overcast) (Fig. 6) days. Note that data points shown on sub-figures of Fig. 3 (a and f) through Fig. 6 (a and f) include additional R_n stations and pan evaporation stations used to obtain ground-based R_n and ground based DAET maps. These stations are not included in Fig. 1 or Table 1 which only has the stations used

Discussion and conclusion

The objective of the current study was to develop and validate a robust and reliable daily ET estimation model for all sky (comprising of cloudy and clear sky) conditions using remote sensing data. Our proposed SatDAET algorithm first estimates the Evaporative Fraction (EF) by utilizing the relationship between NDVI and radiometric surface temperature observed from AVHRR for each day. Then spatio-temporal interpolation and filtering techniques are applied to obtain daily EF values for cloudy

Acknowledgements

This work was supported by a contract from the South Florida Water Management District and accomplished through collaboration among SFWMD, University of Cincinnati, Tufts University, MIT, NOAA/NESDIS/ORA and I. M. Systems Group, Inc. Part of this work was also supported by a grant from the National Research Institute Competitive Grants Program of the United States of America (MASR 2004-00888). Comments on technical implementation from SFWMD, and from USGS scientist Dr. Sumner to refine the

References (23)

BatraN. et al.
Estimation and comparison of evapotranspiration from MODIS and AVHRR sensors for clear sky days over the Southern Great Plains
Remote Sens. Environ.
(2006)
BishtG. et al.
Estimation of net radiation using MODIS (Moderate Resolution Imaging Spectroradiometer) terra data for clear sky days
Remote Sens. Environ.
(2005)
CarlsonT.N. et al.
An interpretation of methodologies for indirect measurement of soil water content
Agric. For. Meteorol.
(1995)
CarlsonT.N. et al.
On the Relationship Between NDVI, Fractional Vegetation Cover and Leaf Area Index
Remote Sensing Environment
(1997)
KustasW.P. et al.
Using satellite remote sensing to extrapolate evapotranspiration estimates in time and space over a semiarid rangeland basin
Remote Sens. Environ.
(1994)
VenturiniV. et al.
Comparison of evaporative fractions estimated from AVHRR and MODIS sensors over South Florida
Remote Sens. Environ.
(2004)
BastiaanssenW.G. et al.
Estimation of surface resistance and Priestley–Taylor — parameter at different scales
BettsA.K. et al.
FIFE surface climate and site-average dataset 1987–89
J. Atmos. Sci.
(1998)
BrasR.
GermanE.R.
Regional evaluation of evapotranspiration in the everglades

HallF.G. et al.

Status of remote sensing algorithms for estimation of land surface state parameters

Remote Sens. Environ.

(1995)

Cited by (95)

Integration of microwave satellite soil moisture products in the contextual surface temperature-vegetation index models for spatially continuous evapotranspiration estimation
2023, ISPRS Journal of Photogrammetry and Remote Sensing
The contextual surface temperature-vegetation index (TVX) models have been widely used for the retrieval of soil moisture (SM) and evapotranspiration (ET). One of the key premises for their application is to determine quantitatively the theoretical boundaries of this contextual TVX space. Usually these theoretical boundaries are determined based on land surface energy balance principle. Although sound in physical mechanism, the complex parameterization involved has hampered its remote sensing (RS) applications. Our recent studies show that ground-based SM observations can be used for a quick and continuous retrieval of these theoretical boundaries. The main disadvantage lies in its reliance on in-situ SM observations. Fortunately the development of microwave RS technology has made it possible to monitor SM at both regional and global scale. Under this background, a practical framework based on the contextual TVX models was proposed in this paper for continuous ET estimation through the combination of microwave-based SM products and MODIS (Moderate Resolution Imaging Spectroradiometer) products. Daily actual ET over the Southern Great Plains of the United States of America was calculated as the product of net radiation and evaporative fraction (EF). The focus of this framework was to investigate how optical and microwave RS can be combined to achieve a quick and continuous retrieval of EF independently of ground observations. Specifically, EF was retrieved from the contextual TVX models, the theoretical boundaries of which were calibrated at the annual scale based on an optimization scheme. Assuming the seasonal variation of the theoretical dry edge follows a cosine function of solar zenith, the essence of this optimization scheme is to search for the optimal amplitude parameter that maximizes the correlation between microwave-based SM observations and soil moisture index retrieved from the contextual TVX models. Two microwave-based SM products including SMAP (Soil Moisture Active Passive) and SMOS (Soil Moisture and Ocean Salinity) were adopted to demonstrate this optimization scheme. The correlation coefficient, root mean square error and bias of EF estimates calibrated by SMAP were 0.700, 0.146 and 0.047, and those achieved by SMOS were 0.663, 0.174 and 0.168, respectively. The optimization scheme has not only made it possible to conduct a continuous monitoring of ET based entirely on RS observations but also achieved the downscaling of microwave satellite SM products. Thus the combination of optical and microwave RS holds great potential for coupled and continuous estimation of SM and ET over large scale.
Triple collocation-based merging of multi-source gridded evapotranspiration data in the Nordic Region
2023, Agricultural and Forest Meteorology
Accurate evapotranspiration (ET) data are required for many hydro-meteorological applications. Compared with the traditional evaluation that requires in-situ measurements, the triple collocation (TC) technique estimates geophysical product errors without the need for ground truth, which is especially suitable over large areas lacking a dense in-situ network. However, violations of the zero-error cross-correlation (ECC) assumption are found to be the dominant sources of impairing the TC robustness. This study presents the first application of a TC-based merging framework that optimally considers ECC to merge multi-source gridded ET products in the Nordic Region during 2003–2018. The ECC estimates of each ET dataset pair calculated by the quadruple collocation approach are used to select the qualified triplets from four products, including FLUXCOM, Global Land Surface Satellite (GLASS), Global Land Evaporation and Amsterdam Model (GLEAM), and Penman-Monteith-Leuning Version 2 (PML-V2). Then the ET merged datasets are generated by weighting TC-based rescaled error variances of the parent datasets through least square merging. Finally, the accuracy of both the parent and the merged datasets are assessed with the Integrated Carbon Observation System (ICOS) flux data in the Nordic Region based on multiple statistical metrics. Results demonstrate that the ECC values provide intuitive evidence for filtering unqualified TC triplets. Both the absolute and relative error variances (signal-to-noise ratio) are considered for ET dataset evaluation. Overall PML-V2 has the best performance among the evaluated four products. Two merged ET datasets with the reference climatology of FLUXCOM outperform all parent products with the lowest errors by using ICOS data as reference among all sites – indicating the feasibility of TC technique for improving ET accuracy in the Nordic Region. This study also analyses the impacts of reference climatology selection on the TC merged results.
Estimation of crop evapotranspiration from MODIS data by combining random forest and trapezoidal models
2022, Agricultural Water Management
Citation Excerpt :
Remote sensing-based methods have been widely used for cropland ET calculation and monitoring from regional to global scale (Jéffersonde et al., 2020; Jiang et al., 2020; Knipper et al., 2020; Liu et al., 2019; Mu et al., 2007, 2011; Xue et al., 2020; Yu et al., 2019), and can be categorized into three groups based on their principles: (1) surface energy balance methods, (2) surface water balance methods, and (3) land surface temperature (LST)/fractional vegetation cover (FVC) feature space methods (Carlson, 2007; Li et al., 2009; Senay, 2008; Su, 1988). Among these methods, LST/FVC feature space methods are advantageous in ET calculation because they avoid complex aerodynamic and land surface resistance parameterization required for other methods (Jiang et al., 2009; Kimball and Running, 2016; Leng et al., 2017). The hypothesis of the LST/FVC feature space methods is that remotely sensed LST and FVC of all clear-sky pixels in an image could be used to generate a two dimensional space, and the envelope of these pixels could constitute a physically meaningful triangle or trapezoid (Carlson, 2007).
Evapotranspiration (ET) is an important parameter for crop growth monitoring and land surface modeling. This paper proposed a new workflow, namely ESVEP-RF, to calculate ET during the crop growing season using MODIS data by combining the advantages of the trapezoidal model and Random Forest (RF) algorithm. In ESVEP-RF, the endmember-based soil and vegetation energy partitioning (ESVEP) model was first used to calculate a series of parameters from MODIS and meteorological inputs, and then all parameters derived from remote sensing data, meteorological data and ESVEP models were used as inputs to the RF algorithm for latent heat flux (LE) calculation. In-situ data of 12 years (2003–2012, 2018 and 2019) from five flux towers located in Nebraska (NE) and Michigan (MI) were used to test the performance of ESVEP-RF, and results showed that ESVEP-RF had great potential to accurately calculate ET when the number of training samples was sufficient and representative. In 2010 and 2011, R² of LE were around 0.8 and RMSE were around 70 W/m², which outperformed original ESVEP model results. This indicated that the RF algorithm could better describe the non-linear correlation between in LST/FVC space endmembers and LE. Among all parameters, LAI, PLEv and R-vw had high contribution with percentage importance of 18.49%, 15.71% and 13.57%, respectively. Furthermore, all samples between 2003 and 2012 collected from the three NE sites were used to train RF models and then calculate LE for both NE and MI sites in 2018 and 2019. In NE sites, RMSE was around 65 W/m² and R² was around 0.8. In MI sites, it was noted that no samples from these sites were included in the training data set, and RMSE was around 70 W/m² and R² was higher than 0.7. These results showed the potential of ESVEP-RF for providing up-to-date ET information.
An operational parameterization scheme of surface temperature-vegetation index contextual model for large-scale temporally continuous evapotranspiration estimation: The case study of contiguous United States
2021, Journal of Hydrology
A recent review has suggested relatively simple contextual models would remain attractive in the estimation of evapotranspiration (ET). The surface temperature-vegetation index feature space model is a typical representative since it provides direct monitoring of ET using simply and effective contextual information. However, the temporally continuous estimation and large-scale application issues are not well identified at previous studies. Under this background, the main motivation for this study is to develop an operational new parameterization scheme within the feature space framework for large-scale temporally continuous ET estimation. Specifically, the spatial domain division strategy was first designed to obtain the optimal grid division size for large-scale ET estimation. The model boundaries of these sub-regions were parameterized using spatial domain based temporal domain information at seasonal scale. The applicability of new method was demonstrated over the entire contiguous United States (CONUS) in the year 2008 using Moderate Resolution Imaging Spectroradiometer (MODIS) and Global Land Surface Satellite (GLASS) products. The comparison between present method and traditional method indicates that the new parameterization scheme produced at different reference spatial domains has a slight accuracy improvement over the 14 FLUXNET sites. In addition, the new method holds unique advantages in its simplicity and continuity, especially superior in the temporally continuous estimation of ET over a large extended region. The validation results of 14 FLUXNET sites indicate the correlation coefficient (r), mean absolute error (MAE), root mean square error (RMSE) the mean bias (B) values for continuous estimation of daily ET were 0.62, 28.430 W/m², 36.133 W/m² and −0.059, respectively. The comparison results with existing ET product also demonstrated the comparable of ET distribution and estimates. Meanwhile, the boundary parameters obtained at seasonal scale also enhanced the physical meanings of feature space model. Consequently, the new parameterization scheme proposed in this study can be regarded as a suitable tool to provide large-scale temporally continuous estimation of ET.
Generation of spatio-temporally continuous evapotranspiration and its components by coupling a two-source energy balance model and a deep neural network over the Heihe River Basin
2021, Journal of Hydrology
Evapotranspiration (ET) and its components of soil evaporation (E) and vegetation transpiration (T), as key variables for the water-energy exchange between the land surface and the atmosphere, are widely used in hydrological and agricultural applications. The land surface temperature based two-source energy balance (TSEB) model can provide high accuracy E, T and ET, which are spatio-temporally discontinuous, whereas the spatio-temporally continuous daily ET is more helpful in water resources management. In this study, to improve the continuity of estimates from the TSEB model, we developed a new combined model coupling the TSEB model and deep neural network (DNN) (TSEB_DNN). First, spatio-temporally continuous reference data was prepared based on the remote sensing and meteorological data as input, and E from soil and T from vegetation were obtained from the TSEB model under clear-sky condition as outputs. Then, the DNN was trained under clear-sky condition to obtain the relationship between E and T estimates from TSEB and reference data. Finally, the trained DNN was driven by the spatio-temporally continuous reference data to obtain spatio-temporally continuous E, T and ET. Compared with the ET estimates from the original TSEB model, the continuity was significantly improved for the TSEB_DNN model. The TSEB_DNN model was well consistent with the in situ measurements and had the overall correlation coefficient (R), root-mean-square-error (RMSE), and bias values of 0.88, 0.88 mm d⁻¹, and 0.37 mm d⁻¹, respectively. The ratio of T/ET estimates from the TSEB_DNN model had high accuracy against in situ measurements with RMSE and bias values of 7.49% and −2.22%, respectively. The combined model and the maps of E, T and ET will help improve water resource management.
Critical comparison of actual evapotranspiration estimates using ground based, remotely sensed, and simulated data in the USA
2021, Agricultural Water Management
Agriculture is a major consumptive user of water through evapotranspiration (ET). Quantification of actual ET (ET_a) is of significant interest to better address water management and improve decision making in the agricultural sector. Many methods to estimate ET are available. However, due to the increasing need of validation of estimates of ET from various techniques, there arises a call for a critical comparison of ET_a estimated from a combination of methods. This research presents a critical comparison of ET_a estimates obtained using ground based (eddy covariance), remotely sensed (ET ensemble), and simulated data (a crop-soil-water balance model, SIMETAW#). The study focuses on experimental sites in Nebraska, Oklahoma and California (USA). Results show a relatively small and consistent variation between the three methods. The results are relatively close giving a reasonable impression of the expected ET_a at a site including variability which is useful to address agricultural practices. The estimates from remote sensing deviate from the other two methods since the method used is relatively new and requires further research to improve measurements and ET estimation. This study offers water managers at the sites new information for reliable, multi-source ET estimation for water resources planning and the study is in principle replicable at different locations.

View all citing articles on Scopus

View full text

A satellite-based Daily Actual Evapotranspiration estimation algorithm over South Florida

Abstract

Introduction

Section snippets

Basic framework

Remote sensing data processing

Results of DAET maps for clear and cloudy days

Discussion and conclusion

Acknowledgements

Remote Sens. Environ.

Remote Sens. Environ.

Agric. For. Meteorol.

Remote Sensing Environment

Remote Sens. Environ.

Remote Sens. Environ.

Estimation of surface resistance and Priestley–Taylor — parameter at different scales

FIFE surface climate and site-average dataset 1987–89

J. Atmos. Sci.

Regional evaluation of evapotranspiration in the everglades

Status of remote sensing algorithms for estimation of land surface state parameters

Remote Sens. Environ.