Global Drought and Flood. Группа авторов
Читать онлайн книгу.
water balance. Remote sensing approaches have been developed to retrieve regional or global scale evapotranspiration in recent decades. As the lack or reduction of evapotranspiration indicates drought, remote sensing of evapotranspiration has been applied to monitoring regional or global droughts in recent years. In this chapter we briefly review ET remote sensing studies, starting with a historical sketch before introducing the Geostationary Operational Environmental Satellites’ (GOES) ET and Drought (GET‐D) product system that is operational at the National Environmental Satellite, Data, and Information Service (NESDIS). The GET‐D system implements the Atmosphere–Land Exchange Inversion (ALEXI) model for estimating regional daily ET from observations of the NESDIS Geostationary Operational Environmental Satellites. The Evaporative Stress Index (ESI) based on ALEXI ET is used for monitoring drought currently for North America. An approach to merging the ESI data into microwave soil moisture observations and land‐surface model soil‐moisture simulations for a blended drought index is presented. The feasibility of using the ALEXI ET estimates from global satellite observations for drought monitoring is discussed.
2.1. INTRODUCTION
Evapotranspiration (ET) is the sum of water evaporated from Earth’s surface, from both land and ocean, and water transpired from vegetation. Thus, ET is commonly referred to as evaporation from the land surface, including soil surface evaporation, the evaporation of water intercepted by vegetation canopy, and the transpiration from vegetation stomata.
The latent heat needed for the evapotranspiration processes and transferred to the atmosphere is one of the most important components of the global or regional energy cycle. Partitioning of the available energy of a surface between latent heat (LE) and sensible heat fluxes can affect atmospheric motions and can influence local and regional weather via temperature and moisture advection and atmospheric motion. Evapotranspiration, or LE, is the largest energy source for the atmosphere and thus it is a critical factor for weather and climate formation (Rabin et al., 1990).
Evapotranspiration is a major component of the global water cycle. The hydrological budget of many land surfaces is the partitioning of precipitation between ET and runoff. The antecedent ET can largely determine the antecedent soil moisture condition, which in turn plays a significant role in the amount of runoff at the watershed scale (Loague & Freeze, 1985). Thus, many hydrological models for the monitoring or forecasting of streamflow require estimates of ET rates.
Both evaporation and transpiration require liquid water to be available for vaporization and the energy to be used to vaporize the water. The availability of the water and the energy varies significantly both spatially and temporally. Thus the spatial and temporal variations of ET could be tremendous. Lack or reduction of ET indicates low water availability or water deficit, i.e., drought. Therefore, observing ET is one approach to monitoring drought. For vegetated surfaces or crop fields, ET rates, especially transpiration rates, are closely related to the need of plants or crops to assimilate carbon for their maintenance and growth (Monteith, 1988). Thus, significant deviation from a potential or optimal ET rate indicates water stress for the plants or crops, i.e., agricultural drought. Monitoring agricultural droughts regionally or globally has significant economic and even political implications.
Local scale evapotranspiration is mostly observed using ground instruments, such as a lysimeter, the Bowen ratio, or eddy covariance tower. Regional and global scale evapotranspiration are estimated from modeling land surface or atmospheric water and energy budgets. Remote sensing approaches have been developed to retrieve regional or global scale evapotranspiration in recent decades (Courault et al., 2005; Kustas & Norman, 1996; Vinukollu et al., 2011). Because of the relatively low expense and continuous spatial coverage, satellite remote sensing of evapotranspiration has been applied to monitoring regional or global droughts in recent years. In this chapter, we give a historical sketch of the ET remote sensing studies and existing ET data products in the next section. Then in section 2.3 we describe how to use the Atmosphere–Land Exchange Inversion (ALEXI) model for operationally estimating ET and monitoring drought with observations in the Geostationary Operational Environmental Satellites’ (GOES) ET and Drought (GET‐D) product system of the National Oceanic and Atmospheric Administration’s (NOAA) National Environmental Satellite, Data, and Information Service (NESDIS). Section 4 presents an example of combining various satellite remote sensing data products, including ET, for monitoring drought. Finally section 5 discusses how satellite observations could provide operational global ET observations for climate studies and various societal applications.
2.2. HISTORICAL SKETCH OF ET REMOTE SENSING STUDIES AND ET DATA PRODUCTS
Recognition of the natural evaporation processes might have started about 500 BCE according to the chronological sketch in Brutsaert (1982), but the search for understanding the evaporation and plant transpiration processes may not have begun until a couple of centuries ago. Early efforts of estimating evaporation or ET rates include empirically relating potential evaporation rate to factors such as near‐surface vapor pressure deficit and a ratio of the monthly daily air temperature over a heat index that depends on the 12‐month mean air temperature (Thornthwaite, 1948). For open water or surfaces without water limitation, Penman (1948) derived the famous Penman equation for estimating potential or optimal evaporation rate from shortwave radiation, vapor pressure deficit, daily mean temperature, and wind speed. Introducing a vegetation canopy or surface conductance into the derivation of the Penman equation, Monteith (1965) developed the widely used Penman–Monteith equation to estimate actual surface ET that depends on net shortwave radiation, ground heat flux, vapor pressure deficit, daily air temperature, wind speed, aerodynamic conductance and canopy conductance. By removing the aerodynamic term and adding an empirical surface related factor alpha to the Penman–Monteith equation, Priestley and Taylor (1972) developed the well‐known Priestley–Taylor equation for approximating ET rate. Many variations of the above equations and other ET observation and estimation approaches have been developed and used for ET rates of different land surface at different spatial and temporal scales (see review in Wang & Dickinson, 2012).
With the emergence of remote sensing technology, many studies have started to estimate ET rates from the remotely sensed, spatially distributed observations of land surface properties in recent decades. Land surface temperature data indicate the state of the land surface and the partitioning of the available energy (the net radiation minus the soil heat flux) into sensible heat and latent heat. Satellite remote sensing is the only technology able to provide radiometric surface temperature observations at the global scale (Kustas & Norman, 1996). Optical satellite sensors can provide information on land surface type and vegetation dynamics, which is required by some advanced algorithms used to estimate ET, such as the Penman–Monteith equation (Monteith, 1965). The effectiveness, efficiency, and economic advantage of obtaining globally spatial distributed input data from satellite remote sensing have led to active international research activities for ET estimation (Sellers et al., 1990). Price (1980) started to use NOAA Advanced Very High Resolution Radiometer (AVHRR) day–night pairs of land surface temperature and daily average climate data with the daily integrated land surface energy balance equation to estimate daily fluxes including ET. Price (1982) further refined his method and obtained ET with reasonable accuracy when comparing estimates based on meteorological data and pan evaporation data. Assuming that the instantaneous differences between radiometric land surface temperature and air temperature is directly related to daily ET, R. D. Jackson et al. (1977) derived a simple relationship for field scale ET estimation. Similar approaches are then widely used for mapping daily ET over large areas from observations of land surface temperature (Allen et al., 2005; Carlson & Ripley, 1997; Courault et al., 1994; Lagouarde, 1991). To estimate actual ET, satellite observed vegetation indices are widely used to compute a reduction factor (e.g., the Priestley–Taylor parameter alpha) from potential ET based on ground meteorological forcing data (e.g., Allen et al., 2005; Kalluri et al., 1998; Neale et al., 2005).
Kustas and Norman (1996) reviewed these approaches and classified them into three general categories: empirical or semiempirical/statistical, physical/analytical, and numerical modeling approaches. More than a dozen ET models and studies of these