Estimating Fluxes on Continental Scales Using Remotely Sensed Data in an Atmospheric–Land Exchange Model

John R. Mecikalski Cooperative Institute for Meteorological Satellite Studies, Space Science and Engineering Center, University of Wisconsin—Madison, Madison, Wisconsin

Search for other papers by John R. Mecikalski in
Current site
Google Scholar
PubMed
Close
,
George R. Diak Cooperative Institute for Meteorological Satellite Studies, Space Science and Engineering Center, University of Wisconsin—Madison, Madison, Wisconsin

Search for other papers by George R. Diak in
Current site
Google Scholar
PubMed
Close
,
Martha C. Anderson Department of Soil Science, University of Wisconsin—Madison, Madison, Wisconsin

Search for other papers by Martha C. Anderson in
Current site
Google Scholar
PubMed
Close
, and
John M. Norman Department of Soil Science, University of Wisconsin—Madison, Madison, Wisconsin

Search for other papers by John M. Norman in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

A simple model of energy exchange between the land surface and the atmospheric boundary layer, driven by input that can be derived primarily through remote sensing, is described and applied over continental scales at a horizontal resolution of 10 km. Surface flux partitioning into sensible and latent heating is guided by time changes in land surface brightness temperatures, which can be measured from a geostationary satellite platform such as the Geostationary Operational Environmental Satellite. Other important inputs, including vegetation cover and type, can be derived using the Normalized Difference Vegetation Index in combination with vegetation and land use information. Previous studies have shown that this model performs well on small spatial scales, in comparison with surface flux measurements acquired during several field experiments. However, because the model requires only a modicum of surface-based measurements and is designed to be computationally efficient, it is particularly well suited for regional- or continental-scale applications. The input data assembly process for regional-scale applications is outlined. Model flux estimates for the central United States are compared with climatological moisture and vegetation patterns, as well as with surface-based flux measurements acquired during the Southern Great Plains (SGP-97) Hydrology Experiment. These comparisons are quite promising.

Corresponding author address: Dr. John R. Mecikalski, CIMSS Space Science and Engineering Center, University of Wisconsin—Madison, 1225 West Dayton St., Madison, WI 53706.

johnm@ssec.wisc.edu

Abstract

A simple model of energy exchange between the land surface and the atmospheric boundary layer, driven by input that can be derived primarily through remote sensing, is described and applied over continental scales at a horizontal resolution of 10 km. Surface flux partitioning into sensible and latent heating is guided by time changes in land surface brightness temperatures, which can be measured from a geostationary satellite platform such as the Geostationary Operational Environmental Satellite. Other important inputs, including vegetation cover and type, can be derived using the Normalized Difference Vegetation Index in combination with vegetation and land use information. Previous studies have shown that this model performs well on small spatial scales, in comparison with surface flux measurements acquired during several field experiments. However, because the model requires only a modicum of surface-based measurements and is designed to be computationally efficient, it is particularly well suited for regional- or continental-scale applications. The input data assembly process for regional-scale applications is outlined. Model flux estimates for the central United States are compared with climatological moisture and vegetation patterns, as well as with surface-based flux measurements acquired during the Southern Great Plains (SGP-97) Hydrology Experiment. These comparisons are quite promising.

Corresponding author address: Dr. John R. Mecikalski, CIMSS Space Science and Engineering Center, University of Wisconsin—Madison, 1225 West Dayton St., Madison, WI 53706.

johnm@ssec.wisc.edu

Introduction

In the past decade, the role of the land surface in modulating the global climate and influencing day-to-day weather events has gained increased recognition. A variety of land surface parameterizations have been developed for models of climate and weather, incorporating varying levels of complexity (Smith et al. 1988; Mahfouf et al. 1995; Wetzel and Boone 1995; Sellers et al. 1996; DeRidder and Schayes 1997; and Rosenzweig and Abramopoulos 1997 are a few current examples). A sizable research effort, the Project for Intercomparison of Land Surface Schemes (PILPS; Henderson-Sellers et al. 1993), has been dedicated to evaluating the relative accuracy of existing parameterizations. An outstanding challenge in PILPS lies in verifying the flux estimates generated by these models over the large spatial scales that they represent.

Our goal here is to detail a robust and computationally tractable method for evaluating the surface energy budget over large geographical regions using input that can be derived primarily through remote sensing. Surface-flux-estimation models developed for localized field studies are typically ill suited for regional-scale application because they require site-dependent tuning (Kustas et al. 1996). Many of these detailed numerical models quickly become computationally prohibitive when applied over tens of thousands of grid cells, and their complexity cannot be justified if the detailed surface information that they require is not available. In contrast, simpler models often rely heavily on surface meteorological data that cannot be obtained with adequate accuracy and/or horizontal resolution over large regions. The potential applications for a reliable regional-scale flux model, however, are a strong motivation for further development. These applications exist in the fields of climate modeling and weather forecasting, hydrology, agriculture, forestry, and many other related disciplines.

This work uses a methodology recently developed by Anderson et al. (1997) that does not suffer from large data requirement and data processing handicaps; only a modicum of supporting data and computer processing time are necessary to make continental-scale flux evaluations. The Atmosphere–Land Exchange Inversion (ALEXI) model has been designed to mitigate many of the problems encountered in earlier land surface flux estimation methods that employ thermal infrared (TIR) measurements of the surface.

Using surface temperature measurements acquired with ground-mounted infrared thermometers, Anderson et al. (1997) found that ALEXI model flux estimates agreed well with measurements obtained during the First International Satellite Land Surface Climatology Project (ISLSCP) Field Experiment (FIFE; Sellers et al. 1988) and the Monsoon ’90 experiment (Kustas et al. 1991). The representative horizontal scale of these flux estimates was on the order of 100 m. Here, we will apply the ALEXI model on continental spatial scales at a resolution of 10 km, using surface temperature measurements acquired with the Geostationary Operational Environmental Satellite (GOES-8) satellite as model input.

The paper proceeds as follows. The basic and unique aspects of the model used for this study are highlighted in section 2, while section 3 details the production of the input datasets required by the model. Section 4 presents model results generated for several case study days. Conclusions and directions for future work are presented in section 5.

The Atmosphere–Land Exchange Inversion Model

Background

The ALEXI model has been described in detail by Anderson et al. (1997). Originally called the Two-Source Time-Integrated Model (TSTIM), this model was an extension of the earlier Two-Source Model (TSM) developed by Norman et al. (1995). Both models assume that surface fluxes arise primarily from two sources: the soil and vegetation covering the surface. The TSM uses a single-time observation of surface brightness temperature and attains sensible heat flux closure with shelter-level measurements of air temperature. In contrast, the ALEXI model uses brightness temperature data at two times and employs a simple model of the temporal change of the energetics of the atmospheric boundary layer (ABL) for flux closure. The benefits gained from these modifications are discussed below.

Four goals have guided the development of the ALEXI model: 1) Means for mitigating the major sources of error in interpreting infrared thermometry have been specifically incorporated into the model. These will be discussed in the following section. 2) The input to the model has been limited to quantities that can be acquired with reasonable accuracy over large geographical regions from remote sensing and/or readily available land use information. The amount of specific ground-based observations required by the model is modest and, where possible, inputs have been parameterized in terms of quantities that can be derived from satellite data. On the other hand, 3) enough generality has been retained in the model so that it can be applied to a wide range of natural surface types without the site-specific empirical adjustments or regional calibrations required by many other similar schemes. Toward this end, simplicity has sometimes been sacrificed for a somewhat more physically based representation. For example, the benefits of using a dual-source (soil plus vegetation) model of the land surface, rather than a single-source model, outweigh the cost of the additional vegetation cover input and associated computations that are required (Kustas et al. 1996). And, 4) the model has been kept computationally simple, so that implementation in near–real time over continental-scale grids containing hundreds of thousands of pixels remains tractable.

Model description

The equation set comprising the ALEXI model is summarized in the appendix and is described in greater detail by Anderson et al. (1997). Figure 1 shows a schematic representation of the model structure; the surface layer component of the model is described in Fig. 1a, and the ABL treatment is shown in Fig. 1b. The symbols are defined in the appendix.

Flux partitioning in the ALEXI model is guided by time changes in surface brightness temperature. The amplitude of the diurnal surface temperature wave has been found to be a good indicator of surface flux partitioning;wetter surfaces warm more slowly and expend more energy in evaporation. Such time-change relationships have been investigated by Idso et al. (1975), Price (1980), Carlson et al. (1981), Wetzel et al. (1984), Diak and Stewart (1989), Diak (1990), and others. Time changes in shelter-level and ABL air temperature also provide information concerning the local surface energy balance. Several relatively simple slab models of the ABL (e.g., Tennekes 1973; Driedonks 1982) relate the rise in temperature and height of the mixed layer to the time-integrated influx of sensible heating from the surface; such models have been used to estimate time integrals of surface fluxes at radiosonde locations and also over regional scales (Diak 1990; Culf 1993; Diak and Whipple 1995; Mecikalski et al. 1997). In the ALEXI model, these two principles are merged, with a simple two-source land surface parameterization providing the link between time changes in surface and air temperature.

The use of time changes of brightness temperatures and the state of the ABL in the ALEXI model was motivated by practical concerns. Working with temporal differences in brightness temperature measurements (in lieu of absolute, single-time-level measurements), causes errors in sensor calibration, atmospheric corrections, and the specification of surface emissivity to be mitigated. Such errors have been detrimental to methods that rely on determination of the absolute surface radiometric temperature (generally incorporating information at a single time), and the use of an absolute surface–air temperature difference to estimate fluxes. In ALEXI, the time-independent component of all resulting biases in the derived surface radiometric temperature drops out in the time difference. Both the surface layer and ABL components of the model run in a time-differential mode; absolute values of brightness temperature and air temperature do not enter into flux determinations.

Soil, canopy, and total energy budgets are balanced within the ALEXI model. These budgets are forced by the net radiation that is partitioned between the soil and vegetative components of the scene; the divergence of net radiation within the canopy (Rn,c) depends on the local fraction of surface vegetation cover. A fixed fraction of the net radiation at the soil surface (Rn,s) is conducted into the submedium, a reasonable assumption during the daytime hours (Choudhury et al. 1987). This simplification of the soil flux term (G) eliminates the need for a time-dependent soil model and the specification of soil thermal and hydraulic properties at each modeling site (information that is not readily available on continental scales), thus greatly speeding computations.

Given an estimate of the fraction of vegetation cover in the scene, fc(ϕ), the ALEXI model extracts the soil and vegetation temperatures (Ts and Tc, respectively) from the satellite-derived ensemble radiometric temperature, TRAD (ϕ). The sensor look angle (ϕ) is retained in both the radiometric temperature and fractional vegetation cover expressions to accommodate off-nadir surface temperature measurements made from satellite platforms (Anderson et al. 1997). Using Ts, Tc, and an initial estimate of the near-surface air temperature (Ta), the surface component of the ALEXI model computes soil and canopy sensible heat fluxes (Hs and Hc). A simple Priestley–Taylor approximation (Priestley and Taylor 1972) then provides a first-guess estimate of canopy transpiration (LEc). This approximation assumes that the canopy is transpiring at its potential rate, which is often an overprediction in semiarid and arid ecosystems, when soil moisture is limiting, or, in general, when plants are stressed or transpiration is reduced for any reason. Soil evaporation (LEs) is then computed as the residual to the system energy budget. If this computation results in a negative soil latent heat flux (i.e., condensation on the soil, which is an unrealistic condition during daytime hours), the Priestley–Taylor estimate of canopy transpiration is reduced until soil latent heating is zero. System energy budget constraints then allow a new estimate of Ta to be computed.

The surface-layer part of the ALEXI model is executed twice, using surface brightness temperatures from GOES acquired at two times during the morning hours, and a corresponding change in near-surface (50 m) air temperature is derived. Given an early morning sounding of the atmospheric temperature profile, a simple slab model of the ABL diagnoses the growth in boundary layer height and the time-integrated input sensible heating required to raise the ABL height to its diagnosed value at the second time (see Fig. 1b). In this application, an energy conservation equation with a very simple diagnostic entrainment parameterization (McNaughton and Spriggs 1986) was used. If a specific form for sensible heating as a function of time is assumed, then the time integral of the sensible heat flux (H) over the observation interval, as well as two instantaneous H values at the first and second observation times, can be computed.

For computational simplicity, a linear rise in sensible heating during the morning, which is usually a good approximation at this time of the day (Tennekes 1973), is prescribed. In practice, the surface and atmospheric components of the model are iterated, altering the near-surface air temperature estimates until convergence in the time-integrated sensible heat fluxes deduced from each model component is achieved.

Enforcing energy closure using the characteristics of the ABL has proven advantageous in estimating the land surface energy balance from remote sensing and/or synoptic data sources (Diak 1990; Diak and Whipple 1993, 1995; Culf 1993; Mecikalski et al. 1997). The primary benefit gained by coupling a boundary layer model with the two-source surface model is that the need for ancillary measurements of Ta is eliminated. An air temperature is computed for a level of 50 m above the canopy height and is used by the model as a boundary condition that is common to both its surface and ABL components. Flux closure for sensible heating is effectively moved from shelter level (2 m) up to the height of the morning boundary layer. Models that rely on measurements of shelter-level air temperatures as an upper boundary condition suffer significantly when applied over large regions because spatial interpolation between surface weather station measurements (with an average spacing of about 100 km) can yield poor and often biased estimates of local shelter-level temperature (Anderson et al. 1997). Additionally, the small vertical distance between such low-level air temperature measurements and the surface makes flux evaluations, which rely on this vertical surface–air temperature gradient, very susceptible to errors in either air or radiometric temperatures.

The ALEXI model requires brightness temperature measurements at two times during the morning hours to obtain a temperature-change signal. Based on sensitivity tests reported by Anderson et al. (1997), these two observation times have been fixed at 1.5 and 5.5 h past local sunrise. On a clear day, sensible heating increases nearly linearly during this time interval, satisfying the linear functional form for H(t) assumed in ALEXI (see the appendix). Furthermore, Anderson et al. (1997) found that this time interval maximizes both the correlation between sensible heating and surface temperature change, and the brightness-temperature change signal itself. Wetzel et al. (1984) drew similar conclusions, further noting that midmorning meteorological conditions are well suited for this type of investigation because temperature advection is typically at a minimum and skies are most likely to be clear.

Input data production

Overview

To demonstrate an application of the ALEXI model over regional scales, we have selected a modeling region in the central United States: the area bounded by 30°–48°N, and 85°–105°W (see Fig. 2). This region was chosen because it contains strong gradients in climate, annual precipitation, land cover, and land use practices. These gradients should be evident in the output flux fields generated by the model, and the widely varying surface conditions provide a challenging test for the algorithm.

The climatological east–west gradients in the surface energy balance across the central Great Plains are strong due to the westward decrease in mean annual precipitation, and the resultant sharp changes in both agricultural practices and the character of the native flora. Specifically, annual precipitation in the far-southeast portion of the domain may exceed 150 cm yr−1, while locations along the Front Range of the Rocky Mountains receive only 35–40 cm yr−1. Native vegetation varies from wet-mesic to mesic forests in the east to closed canopy coniferous forests in the far north, tall grass prairie across much of the Corn Belt region extending from the western Great Lakes into the central Great Plains (which today is heavily cultivated), and finally to short grass ecosystems and near-desert conditions in the far west along the Front Range of the Rocky Mountain caldera (dominated by a broken canopy of cactus and short grasses). Cultivated crops in the region are primarily corn and soybean across the Midwest and eastern Great Plains; winter wheat in the western Great Plains; and mixed cotton, soybean, and corn into the middle and lower Mississippi Valley.

The field of view of the GOES-8 satellite encompasses this modeling domain with a maximum off-nadir look angle of less than 55°. The domain has been divided into 10 km × 10 km grid cells, giving 223 cells east–west across the domain and 201 in the meridional direction for a total of 44 823 cells. The choice of the 10-km grid cell size was motivated by several considerations. First, it allows for some averaging between adjacent GOES thermal pixels, which have a resolution of approximately 4 km at nadir. Second, 10-km resolution approximates the “flux footprint” on the earth’s surface contributing to the growth of the ABL through the midmorning observation time. This surface flux footprint is on the order of tens of kilometers at 1200 UTC (the time of the first daily radiosonde launch) and approaches about 100–200 km by 0000 UTC (the second daily launch); as the day progresses, the ABL grows in response to the daily time integral of the sensible heat flux and integrates over a larger geographical area. Thus, 10 km represents a midpoint between the 4-km thermal pixel size and the 20–50-km midmorning radiosonde surface flux footprint. Finally, flux-scaling studies indicate that flux aggregation errors are minimized at approximately the 10-km scale. Divakarla (1997) compared flux estimates generated from 30-m scale inputs (acquired with the Landsat Thematic Mapper—TM) and from 1-km scale inputs from the FIFE site. The 30-m TM fluxes aggregated to 1 km differed significantly from the 1-km fluxes derived from the 1-km inputs. However, aggregation of both 30-m and 1-km fluxes to a 15-km scale yielded similar results. This suggests that surface heterogeneities influencing flux estimates at the FIFE site tend to occur on small (hundreds of meters) scales but become averaged out over larger (∼10 km) spatial scales. Avissar and Schmidt (1998) used large-eddy simulations to demonstrate that even with very low wind speeds, the average daytime flux aggregation scale of the ABL is 5–10 km, and significantly larger for larger wind speeds.

In practice, the ALEXI model is executed at each 10-km grid cell across the domain, subject to data availability and quality restrictions. Table 1 presents a list of all input quantities that are allowed to vary over the ALEXI model domain. These data fields must be assembled from a variety of satellite imagery and surface weather station reports and interpolated to a grid at the resolution of the model application. The following subsections describe this data assembly process.

Surface brightness temperatures

Surface brightness temperatures have been obtained from the GOES-8 Imager 10.8-μm channel, which is the cleanest channel in the atmospheric infrared window. Brightness temperatures observed at the satellite were corrected for atmospheric effects using the single-channel algorithm of Price (1983). The atmospheric parameters (vertical profiles of temperature and humidity) required by this algorithm were obtained from the analysis component of the CRAS [Cooperative Institute for Meteorological Satellite Studies (CIMSS) Regional Assimilation System; Diak et al. 1992; Wu et al. 1995] using radiosonde and surface data at 1200 UTC. Surface brightness temperatures at 1.5 and 5.5 h past local sunrise were computed at each cell from satellite imagery retrieved at hourly intervals. Because local sunrise varies across the model domain, this method required an interpolation in time between the hourly GOES images.

The hourly satellite images were subjected to a simple cloud-screening procedure. Grid cells must be clear at both observation times to avoid corrupted surface temperatures. Our assumption of linearity in sensible heating with time requires that cells also be clear throughout the time interval between observations. The cloud-detection test employed therefore enforces a monotonic increase in brightness temperatures during the morning sampling interval; any substantial decrease in brightness temperature during this interval is interpreted as cloud contamination. Cells identified as cloudy were not processed through the ALEXI model and appear blanked in output fields.

Cloud-cleared and atmospherically corrected brightness temperatures (TB) are converted to TRAD using
TBϕϕTRADϕnϕTSKYn1/n
Here, ε(ϕ) is the directional thermal emissivity of the surface at a view angle ϕ, TSKY is the hemispherical sky brightness temperature, and n is the power in the Stefan–Boltzmann equation appropriate for the wavelength band of the sensor (typically n is set to 4 for the 10–12-μm band; Becker and Li 1990). A semiempirical expression for ε(ϕ) as a function of nominal soil and leaf thermal emissivities (εsoil and εleaf respectively, fixed at 0.94 and 0.97) and fractional vegetation cover was developed using the detailed Cupid soil–plant–atmosphere model (Norman et al. 1990).

Vegetation and land surface characteristics

The definitions of other surface characteristics required by the ALEXI model are drawn from two regional data products developed and distributed by the United States Geological Survey (USGS).

The USGS provides a biweekly composite Normalized Difference Vegetation Index (NDVI) product in near–real time, created from data collected by the Advanced Very High Resolution Radiometer (AVHRR) satellite at 1-km resolution (Eidenshink 1992). These NDVI data have been used to create fields of fractional vegetation cover for the ALEXI model.

The USGS also has compiled a prototype land surface classification for the conterminous United States, based on AVHRR data from 1990, also at a 1-km spatial resolution (U.S. Geological Survey 1995). The primary classification system in this dataset consists of 159 detailed land cover classes, but also included is a coarser 20-bin classification similar to that used in the original Simple Biosphere Model (SiB1; Dorman and Sellers 1989). These 20 land cover classes (Fig. 2) have been used to specify at each grid cell a typical vegetation leaf size, and, in conjunction with our fractional vegetation cover estimate, a surface albedo (described below) and a characteristic canopy height used to estimate surface roughness and displacement height (see Table 2). Scaling from the 1-km resolution of the land cover classification to the 10-km resolution of the model grid was accomplished by identifying the class that occurred with the highest frequency within each 10-km model grid cell.

The 1-km NDVI data were scaled to 10-km resolution by simple area averaging. To estimate fractional vegetation cover from NDVI, a simple procedure described by Carlson et al. (1995) was adapted. First, NDVI is normalized between values associated with zero cover (bare soil, NDVImin) and full cover (NDVImax). In this application, a bare-soil NDVI was defined as the minimum NDVI value observed in the ALEXI domain between mid-March and early August (excluding the lower 5% tail of the distribution). A maximum NDVI was defined in an analogous way (excluding the upper 5% tail). All grid cells classified as water were excluded in the determination of these limits. The NDVI values at a particular location and date were then normalized according to
i1520-0450-38-9-1352-e2
These scaled NDVI* values are less sensitive to errors in atmospheric correction and effects due to varying view angle (Carlson et al. 1995). Fractional vegetation cover in each model grid cell was then estimated as
fc2
(Choudhury et al. 1994; Carlson et al. 1995; Carlson and Ripley 1997; Gillies et al. 1997).
Estimates of the local roughness length for momentum transport (zm) and surface displacement height (d) are needed to compute the aerodynamic resistance to turbulent transport (Ra). Goudriaan (1977) provides some useful parameterizations for zm and d in terms of the vegetation canopy height, fractional vegetation cover, and characteristic leaf size. Here, seasonal canopy heights were estimated by combining information from the land cover classification and fractional cover fields described above. An annual minimum and maximum height (hmin and hmax, respectively) was assigned to each of the 20 land cover classes (see Table 2), and an estimate of the height at the current time was obtained by scaling between these limits according to the current cover fraction in each cell:
hcfchmaxfchmin
Characteristic leaf size has been assigned by land cover class (Table 2) and at this point is not varied by the time in the growing season. Anderson et al. (1997), however, have shown that ALEXI flux results are insensitive to variations in leaf size.

A good estimate of the vegetation greenness fraction fg is important for accurate partitioning of the energy budget because it accounts for senescent vegetation that contributes to the canopy sensible heat flux but not to the latent heat flux. For the case study days evaluated in this work, most vegetation was newly emerged; therefore fg is held constant at unity across the model domain in these examples. Later in the growing season this generally would not be the case, and fg would need to be determined from the seasonal evolution of fc.

The hemispherical albedo (A) of a surface containing a sparse canopy depends on the amount of vegetation present in the scene and on the reflectivity of the soil and vegetation scene components. Monteith and Unsworth (1990) summarize an analytical approximation for A using estimates of vegetation cover, leaf absorptivity (αl), and soil reflectivity (ρs) in the visible and near-infrared wavebands. Goudriaan (1977) showed that this approximation works well for solar zenith angles of less than 65°. In this application, soil reflectivity is held constant across the domain, at 0.08 and 0.18 in the visible and near-infrared, respectively. Leaf absorptivities (used in the Monteith and Unsworth scheme for A) have been assigned by land cover class, as tabulated in Table 2. Values for classes 1–12 are those used by Sellers et al. (1996) in the SiB2 land surface classification. Classes 14–20 represent mixed land cover types and have been assigned appropriately averaged absorptivity values.

Surface and upper-air meteorological data

Two important inputs to the ALEXI model cannot be derived with adequate accuracy from satellite observations: near-surface wind speed (V, used in the transport resistance formulations) and atmospheric lapse rate (∂θ/z, used in the ABL model component). These meteorological input fields must be obtained through spatial analyses of observations acquired from the surface synoptic and radiosonde networks. They are therefore of lower spatial resolution than other input fields (∼100–250 km, which is the typical spacing between radiosonde stations). Near-surface air temperature, vapor pressure (ea) fields [used in the computation of the downwelling longwave component of net radiation—Eq. (8) below], and atmospheric pressure (p) fields have been similarly analyzed, but the model flux estimates are not strongly sensitive to these inputs.

Hourly surface data from the Automated Surface Observation System (ASOS) and Automated Weather Observation System (AWOS) were obtained from the University of Wisconsin Space Science and Engineering Center, where they are archived using the Man–Computer Interactive Data Access System (McIDAS; Suomi et al. 1983). McIDAS performs rudimentary quality controls on the data upon their ingestion and archiving.

The surface meteorological data required for ALEXI were objectively analyzed to the 10-km grid using a simple Cressman type successive correction scheme (Cressman 1959). Hourly temperature, vapor pressure, and surface pressure grids were interpolated in time to produce fields corresponding to 1.5 and 5.5 h past local sunrise. Wind speeds were averaged in time around these two sunrise-based times. This averaging has proven beneficial in mitigating the inherently large temporal variability observed in anemometer-level measurements of wind speed, resulting in more accurate surface flux evaluations. Sensitivity tests with ALEXI (Anderson et al. 1997) show that 25% errors in wind speed yield roughly 12% variations in sensible heating.

Upper-air (radiosonde) data at mandatory and significant levels were analyzed to the three-dimensional grid of the CRAS, with a horizontal resolution of approximately 80 km. For this work, 40 vertical levels (25 of them below the 60-kPa level in the atmosphere) were used in the analysis to preserve a high level of vertical detail in the ABL. Subsequently, these vertical profiles were interpolated to the 10-km ALEXI grid and used to produce the required ABL lapse rates.

Net radiation

Net radiation (Rn) is defined as the difference between downwelling and upwelling radiation (all-wave), evaluated at a specific surface. In terrestrial applications, net radiation is partitioned into shortwave (incident and reflected solar radiation: Sd and Su) and longwave (thermal emission from the atmosphere and the surface: Ld and Lu) components:
RnSdSuLdLu
The ALEXI model requires estimates of each of these four radiation components at each cell in the model grid.
Since the ALEXI model can operate only under clear-sky conditions, clear-air models provide the downwelling shortwave and longwave radiation components. The incident solar radiation at the surface (Sd, W m−2) is calculated at the two required times using a simple clear-air model, a component of a system used to estimate solar radiation at the surface from GOES data (Diak and Gautier 1983; Diak et al. 1996). This model modifies the solar input at the top of the atmosphere, taking into account the effects of water vapor absorption, ozone absorption, and Rayleigh and aerosol scattering in the atmosphere to estimate the surface incident solar flux. Results for the clear atmosphere are within about 5% of surface-based measurements. Upwelling (reflected) shortwave radiation (Su, W m−2) is estimated as a function of the local surface albedo and downwelling (incident) shortwave radiation, where
SuASd
The estimation of A from fc information has been described in section 3c above.
Downwelling (atmospheric) longwave radiation (Ld, W m−2) is estimated using the temperature (K) and vapor pressure (kPa) of the lower atmosphere in an empirical expression developed by Lettau and Lettau (1978),
Lde1/2aσT4a
where σ is the Stefan–Boltzmann constant (W m−2 K−4). This model was the most accurate of several simple atmospheric longwave parameterizations evaluated by Redmond (1980).
The ALEXI model extracts the soil and canopy temperatures from the surface radiometric temperature and uses them to compute an upwelling longwave radiation component (Lu, W m−2). Thermal radiation from the soil (Ls) and from the canopy (Lc) will contribute to Lu in proportion to the coefficient of diffuse light transmission through the canopy, τc:
LuτcLcτcLs
Campbell and Norman (1997) give expressions for computing τc given estimates of leaf area index and leaf absorptivity (Table 2) for a variety of leaf angle distributions.

Model results

The ALEXI model solves the equations in the appendix iteratively, generating instantaneous flux estimates of sensible heat, latent heat, soil heat flux, and net radiation at 1.5 and 5.5 h after local sunrise. Because fluxes near sunrise tend to be of low magnitude, attention will be concentrated on the flux distributions generated at the latter time.

Models of large-scale surface fluxes are, in principle, difficult to validate, and this traditionally has led to a certain amount of skepticism concerning the reliability and utility of such models. We therefore take a several-pronged approach toward assessing the quality of the ALEXI flux estimates. Qualitative comparisons can be made with climatological patterns of land cover and moisture. Quantitative comparisons with surface flux measurements acquired during the joint USDA–NASA Southern Great Plains (SGP-97) Hydrology Experiment are also presented. These comparisons provide evidence for the efficacy of this simple flux estimation approach.

Overview of case study days

Two case study days have been selected to demonstrate the utility of the ALEXI model in estimating regional-scale fluxes. The morning of 12 June 1995 was predominantly cloud free over much of the central United States, providing a good opportunity to study regional variations in the flux patterns generated by ALEXI. The day 2 July 1997 was designated as a “Golden Day” during the SGP-97 experiment. Skies were clear over much of the experiment site for most of this day, and therefore it was targeted for intensive surface- and aircraft-based flux evaluations. In the following discussion, these two case study days will be referred to as JUN95 and JUL97, respectively.

12 June 1995

Input fields

Two of the most influential input fields in the ALEXI flux partitioning algorithm are fc and the time change in TRADTRAD), shown for JUN95 in Figs. 3 and 4, respectively. Pixels that appear blanked (white) in the temperature change map failed the cloud-screening test described in section 3b above.

The Corn Belt, stretching from eastern North Dakota and Nebraska through Iowa to central Indiana, exhibits very low vegetation cover on this day. These agricultural regions are planted in May; by June, crops are still in an early stage of development and do not significantly obscure the soil. The southwestern sector of the domain also has low cover, particularly in western Texas and into New Mexico. The steep cover gradient across the southern Great Plains in Kansas, Oklahoma, and northern Texas is associated with a well-known climatological gradient in rainfall distribution. Forested regions in northern Wisconsin and Minnesota and surrounding the Mississippi River Basin are clearly demarcated as areas of high vegetation cover.

Figure 4 shows the surface brightness temperature changes measured by GOES for JUN95. Although the ΔTRAD and fc data fields were derived from completely independent satellites and wavebands, they show remarkably similar spatial features; regions with less cover are associated with large temperature changes. The nearly bare soil in the Corn Belt heats up rapidly during the morning hours. Forests and other well-vegetated regions remain cooler. The strong discontinuity in surface temperature change down the center of the Texas panhandle follows the Caprock Escarpment, separating the High Plains in the west from the lowland Rolling Plains in the east. The Caprock traces the sharp climatological gradients in rainfall and vegetation amounts that characterize the region.

Flux output fields

Modeled fields of sensible and latent (LE) heating at 5.5 h after local sunrise are shown in Figs. 5 and 4, respectively, for JUN95. Energy conservation constrains these fields to mirror each other, given the relatively small magnitude of the soil heat flux term. Land surface patches that are assigned high sensible heating by the ALEXI algorithm have correspondingly low evaporation rates.

Patterns in partitioning between sensible and latent heat fluxes strongly resemble features in the temperature change and cover fraction fields in Figs. 4 and 3. Regions of sparse cover, showing large temperature excursions in Texas and New Mexico, have been assigned low values for latent heating. Across the Corn Belt, the partially bare soils also show low evaporation rates on this June day. In adjacent areas of higher vegetation cover, high transpiration rates and larger roughness lengths (i.e., higher efficiency of turbulent transport) are keeping the surface relatively cool.

An exception to this trend can be found in northern Wisconsin and the Upper Peninsula of Michigan, a heavily forested region exhibiting moderate temperature changes that has been assigned high sensible heating rates. This northern regime is dominated by coniferous forests that have characteristically lower transpiration rates in comparison with deciduous vegetation (Jarvis et al. 1976; Saugier et al. 1997). Higher Bowen ratios and radiometric temperature changes are thus to be expected in these areas.

Values of soil heat flux (not shown) range from approximately 50 to 200 W m−2 over the domain. Larger values of G are associated with regions of low vegetation cover, where there is less interception of downwelling radiation by the canopy, while low values of G are characteristic of forested regions. The field of net radiation values (also not shown) has relative maxima where the surface albedo is low and the rise in surface temperature is relatively small (i.e., net solar radiation is relatively high and outgoing longwave radiation from the surface is relatively low), predominantly in areas with high fractional vegetation cover.

2 July 1997

Input fields

While sky conditions over the SGP-97 experiment domain were mostly clear on the JUL97 case study day, regions elsewhere in the central United States were partially overcast and therefore not processed by ALEXI. Much of the SGP-97 site was under the influence of high pressure, both at the surface and aloft. Weather conditions were characterized by very warm temperatures (35°–40°C) and low relative humidities (<30% by midday).

Input fields of vegetation cover and surface temperature change for JUL97 are shown in Figs. 6 and 4, respectively. Figure 6 shows that crops in the Corn Belt regions in eastern Nebraska, Iowa, Illinois, and Indiana are clearly in a more advanced stage of development than they were on the JUN95 case study day (Fig. 3). Forest canopies in Wisconsin, Minnesota, and along the Mississippi River Valley have also filled in. The radiometric surface temperature changes in these regions are correspondingly lower (compare the two plots in Fig. 4).

Small decreases in cover between JUN95 and JUL97 occur in the northwestern corner of the domain, southward through central Colorado. These decreases may be due to the harvesting of winter wheat, which is grown extensively in that region. The largest surface temperature rises in the domain occur over such regions in eastern Colorado. West of the Caprock in the Texas panhandle, temperature changes remain high, but with reduced amplitude compared to those observed on JUN95.

Flux output fields

Fields of sensible and latent heating for JUL97 are shown in Figs. 7 and 4. Again, these fields are essentially mirror images of one another, with a spatial distribution of Bowen ratios that mimics patterns in the cover and temperature change fields. On this day there is a general increase in evapotranspiration from west to east across the domain, consistent with the growth of new, green vegetation in the east.

Very small evaporative fluxes are evaluated for eastern Colorado where large surface temperature changes were measured. Nebraska was also diagnosed with low latent heating; in this case, sensible heating estimates were enhanced by unusually high morning wind speeds in the northern part of the state. Because latent heat is obtained as the residual to the surface energy budget in the ALEXI model, high wind speeds (generating strong turbulent mixing) enhance sensible heat estimates, thereby decreasing latent heating. This is a reasonable reaction: high winds and high vapor pressure deficits close stomata, causing transpiration to decrease.

The portions of the Corn Belt that passed cloud-screening tests show higher latent heat fluxes on JUL97 than on the JUN95 case study day, presumably due to the advanced stage of maturity (higher vegetation amounts) of crops in these regions. The Caprock again demarcates a transition from low evaporation in the High Plains to the west to higher rates in the Rolling Plains in the east. High evaporation rates are diagnosed for the forested areas along the Mississippi River Basin.

Model validation: Comparisons with surface-based flux measurements

Quantitative comparisons can be made between ALEXI flux estimates for the JUL97 case study day and surface flux measurements acquired during the SGP-97 Hydrology Experiment. Flux measurements during this experiment were primarily clustered around three key sites in Oklahoma: the Agricultural Research Service facilities in El Reno and in the Little Washita watershed southwest of Chickasha, and the ARM CART central facility near Lamont (ARM CART is the Cloud and Radiation Testbed site for the ongoing Atmospheric Radiation Measurement experiment, conducted by the U.S. Department of Energy; Stokes and Schwarz 1994).

Figures 8a–c show comparisons between flux measurements made at the ARM CART, El Reno, and Little Washita sites and ALEXI model predictions for the corresponding grid cell. Where possible, several measurements at a given site have been plotted to demonstrate both the spatial variability in local surface conditions and variability between two different measurement techniques. Fluxes at the ARM CART and El Reno sites were measured with Bowen ratio and eddy covariance systems; the Little Washita data were acquired with an eddy covariance system. The latent heat measurements from all eddy covariance systems have been adjusted to enforce flux closure.

Eddy covariance measurements of available energy (H plus LE) are consistently less than the available energy determined from measurements of net radiation and soil heat conduction fluxes. Because the ALEXI algorithm maintains conservation of energy and the micrometeorological flux measurements do not, the flux measurements have been adjusted to maintain energy conservation, and to provide the most realistic comparison of model and measurements. In SGP-97, great effort was made to calibrate net radiation measurements so we believe them to be the most accurate; thus, we have adjusted the sensible and latent heat fluxes so their sums match the net radiation minus the soil heat flux while maintaining a constant Bowen ratio.

In general, agreement between the modeled and measured fluxes is good; the measurements corroborate the low Bowen ratios diagnosed by the ALEXI model for western Oklahoma on JUL97. ALEXI underestimates (overestimates) latent (soil heating) at the ARM CART site by about 100 W m−2 (Fig. 8a). Because the sensible heating estimate is good, this error most likely occurred because the vegetation cover estimate for this cell (fc = 0.33) was not representative of the sites where the fluxes measurements were taken. If fc is too low, the modeled net radiation at the soil surface will be too high and G will be overestimated. Latent heat is solved as the residual to the surface energy budget, so it must accommodate for this overprediction.

Figure 8b demonstrates the level of variability that can occur among flux measurements taken in close proximity by different measurement systems. Micrometeorological flux measurements typically have uncertainties of about ±20% (see discussion on p. 280 of Norman et al. 1995). In this case, further variability is introduced by heterogeneities in surface conditions across the experiment site. As previously noted, the ALEXI model estimates time-integrated fluxes over the observation interval (here, 4 h), as well as the instantaneous fluxes at individual observation times. A comparison for JUL97 of integrated latent heat fluxes from ALEXI versus those measured at the surface flux stations included in Fig. 8 is given in Table 3.

Conclusions and future research

Using remotely sensed satellite data in combination with routinely available surface and upper-air observations and a modicum of land surface data, the ALEXI model diagnoses the main components of the surface energy balance over large regions. All satellite and atmospheric data comprising the inputs for ALEXI are available in real time, with information on vegetation and land surface characteristics available either biweekly or parameterized as monthly averages. The AVHRR NDVI data used to derive vegetation cover fractions are made available by the U.S. Geological Survey Earth Resources Observation Systems (EROS) Data Center on a weekly and biweekly schedule. Hence, the ALEXI model can be a powerful tool for diagnosing regional fluxes operationally, thus making daily assessments of the surface energy balance over large regions quite feasible. The upcoming launch of the Moderate-Resolution Imaging Spectrometer (MODIS) instrument on the Earth Observing System (EOS) AM-1 platform will provide improved vegetation products for use in models such as ALEXI.

Validating the accuracy of regional surface flux estimates is challenging, but we have shown that both point-flux measurements, together with latent heat inferences obtained via other independent techniques, show reasonable correspondence to the ALEXI model results. As demonstrated, the fluxes produced by the ALEXI model and regional differences in fluxes between the two case study days agree with climatological patterns in vegetation and rainfall.

As with many energy budget assessments that use remotely sensed inputs, the model should not be expected to perform well when the surface energy budget is affected by local circulations (e.g., sea breezes or large values of temperature advection), when ABL warming and the surface brightness temperature response are not necessarily dominated by the surface energy balance. In the future, the ALEXI model may be able to correct for some of these effects through the use of atmospheric predictions from the forecast component of the CRAS.

Regions possessing even a low amount of cloudiness through the morning hours cannot be evaluated via this technique; satellite TIR brightness temperatures quickly become contaminated by clouds that act to cool the scene temperature and yield unrealistic flux estimates. Better cloud detection algorithms (e.g., Hayden et al. 1996, 1998) would make the ALEXI model more robust.

Future work will involve disaggregating the 10-km ALEXI flux estimates to smaller spatial scales using higher-resolution vegetation indices and brightness temperature measurements that are now available from Landsat and AVHRR, as well as similar measurements that will be available in the near future from EOS instruments. A soil moisture inversion algorithm (similar to Kustas et al. 1998), which estimates near-surface soil moisture as a function of soil evaporation, has been incorporated into ALEXI and is now undergoing testing. Soil moisture estimates from this algorithm will be compared with microwave soil moisture estimates made during the SGP-97 program, with the hope of finding synergistic relationships between the two types of soil moisture estimation techniques.

Last, we are in the process of combining brightness temperature measurements from satellites and near-surface air temperature measurements (shown to be of value for estimating the surface energy balance and soil moisture: Mahfouf 1991; Mecikalski et al. 1997) with microwave estimates of near-surface soil moisture into a version of the ALEXI model that is based in the principles of statistical estimation (Diak et al. 1995). The goals of this work are to identify and merge the strengths of these various observing systems, to produce an improved surface flux/near-surface soil moisture product, and to develop a system that is flexible to the incorporation of new data types.

Acknowledgments

Funding for this work was provided by NASA Grants NAGW-4138 and NAG5-3493, and by USDA Cooperative Agreement 58-1270-7-008, with the assistance of the University of Wisconsin Agricultural Experimental Station. The authors would like to thank the individuals who collected and supplied the surface flux measurements used in this work, including Tracy Twine (University of Wisconsin–Madison), Paul Houser (University of Arizona, Tucson, AZ), Pat Starks (USDA/ARS, El Reno, OK), William Kustas (USDA/ARS, Beltsville, MD), John Prueger (USDA/ARS, Ames, IA) and Tilden Meyers (NOAA/ATDL, Oak Ridge, TN). These measurements were collected during the Southern Great Plains Hydrology Experiment, funded by NASA and the USDA. Thanks to T. Carlson for his constructive criticism of an earlier version of this document. We also very much thank W. Kustas for his review.

REFERENCES

  • Anderson, M. C., J. M. Norman, G. R. Diak, W. P. Kustas, and J. R. Mecikalski, 1997: A two-source time-integrated model for estimating surface fluxes using thermal infrared remote sensing. Remote Sens. Environ.,60, 195–216.

  • Avissar, R., and T. Schmidt, 1998: An evaluation of the scale at which ground-surface heat flux patchiness affects the convective boundary layer using large-eddy simulations. J. Atmos. Sci.,55, 2666–2689.

  • Becker, F., and Z.-L. Li, 1990: Temperature-independent spectral indices in thermal infrared bands. Remote Sens. Environ.,32, 17–33.

  • Campbell, G. S., and J. M. Norman, 1998: Introduction to Environmental Biophysics. Springer-Verlag, 286 pp.

  • Carlson, T. N., and D. A. J. Ripley, 1997: On the relationship between fractional vegetation cover, leaf area index, and NDVI. Remote Sens. Environ.,62, 241–252.

  • ——, J. K. Dodd, S. G. Benjamin, and J. N. Cooper, 1981: Satellite estimation of the surface energy balance, moisture availability, and thermal inertia. J. Appl. Meteor.,20, 67–87.

  • ——, W. J. Capehart, and R. R. Gillies, 1995: A new look at the simplified method for remote sensing of daily evapotranspiration. Remote Sens. Environ.,54, 161–167.

  • Choudhury, B. J., S. B. Idso, and R. J. Reginato, 1987: Analysis of an empirical model for soil heat flux under a growing wheat crop for estimating evaporation by an infrared-temperature-based energy balance equation. Agric. For. Meteor.,39, 283–297.

  • ——, N. U. Ahmed, S. B. Idso, R. J. Reginato, and C. S. T. Daughtry, 1994: Relations between evaporation coefficients and vegetation indices studied by model simulations. Remote Sens. Environ.,50, 1–17.

  • Cressman, G., 1959: An operational objective analysis system. Mon. Wea. Rev.,87, 367–374.

  • Culf, A. D., 1993: The potential for estimating regional sensible heat flux from convective boundary layer growth. J. Hydrol.,146, 235–244.

  • DeRidder, K., and G. Schayes, 1997: The IAGL land surface model. J. Appl. Meteor.,36, 167–183.

  • Diak, G. R., 1990: Evaluation of heat flux, moisture flux and aerodynamic roughness at the land surface from knowledge of the PBL height and satellite-derived skin temperatures. Agric. For. Meteor.,52, 181–198.

  • ——, and C. Gautier, 1983: Improvements to a simple physical model for estimating insolation from GOES data. J. Climate Appl. Meteor.,22, 505–508.

  • ——, and T. R. Stewart, 1989: Assessment of surface turbulent fluxes using geostationary satellite surface skin temperatures and a mixed layer planetary boundary layer scheme. J. Geophys. Res.,94, 6357–6373.

  • ——, and M. S. Whipple, 1993: Improvements to models and methods for evaluating the land surface energy balance and “effective” roughness using radiosonde reports and satellite-measured“skin” temperatures. Agric. For. Meteor.,63, 189–218.

  • ——, and ——, 1995: A note on estimating surface sensible heat fluxes using surface temperatures measured from a geostationary satellite during FIFE-1989. J. Geophys. Res.,100, 25 453–25 461.

  • ——, D. Kim, M. S. Whipple, and X. Wu, 1992: Preparing for the AMSU. Bull. Amer. Meteor. Soc.,73, 1971–1984.

  • ——, R. M. Rabin, K. P. Gallo, and C. M. Neale, 1995: Regional-scale comparisons of vegetation and soil wetness with surface energy budget properties from satellite and in-situ observations. Remote Sens. Rev.,12, 355–382.

  • ——, W. L. Bland, and J. R. Mecikalski, 1996: A note on first estimates of surface insolation from GOES-8 visible satellite data. Agric. For. Meteor.,82, 219–226.

  • Divakarla, M., 1997: Estimating spatial distributed surface fluxes from satellite data, in-situ measurements and the Cupid model using GIS (energy flux, carbon dioxide flux, biospheric models). Ph.D. dissertation, University of Wisconsin, 199 pp. [Available from University Microfilm, 305 N. Zeeb Rd., Ann Arbor, MI 48106; AAC 9724097.].

  • Dorman, J. L., and P. J. Sellers, 1989: A global climatology of albedo, roughness length and stomatal resistance for atmospheric general circulation models as represented by the Simple Biosphere Model (SiB). J. Appl. Meteor.,28, 833–855.

  • Driedonks, A. G. M., 1982: Models and observations of the growth of the atmospheric boundary layer. Bound.-Layer Meteor.,23, 283–306.

  • Eidenshink, J. C., 1992: The 1990 conterminous U.S. AVHRR data set. Photogr. Eng. Remote Sens.,58, 809–813.

  • Gillies, R. R., J. Cui, T. N. Carlson, W. P. Kustas, and K. S. Humes, 1997: A verification of the “triangle” method for obtaining surface soil water content and energy fluxes from remote measurements of the Normalized Difference Vegetation Index (NDVI) and surface radiant temperature. Int. J. Remote Sens.,18, 3145–3166.

  • Goudriaan, J., 1977: Crop Micrometeorology: A Simulation Study. Center for Agricultural Publishing and Documentation, 249 pp.

  • Hayden, C. M., G. S. Wade, and T. J. Schmit, 1996: Derived product imagery from GOES-8. J. Appl. Meteor.,35, 153–162.

  • ——, T. J. Schmit, and A. J. Schreiner, 1998: The cloud clearing algorithm for GOES product processing. NOAA/NESDIS Tech. Rep., 19 pp. [Available from A. Schreiner, University of Wisconsin—Madison, Madison, WI 53706.].

  • Henderson-Sellers, A., Z.-L. Yang, and R. E. Dickinson, 1993: The Project for Intercomparison of Land-surface Parameterization Schemes. Bull. Amer. Meteor. Soc.,74, 1335–1349.

  • Idso, S. B., T. J. Schmugge, R. D. Jackson, and R. J. Reginato, 1975:The utility of surface temperature measurements for the remote sensing of surface soil water status. J. Geophys. Res.,80, 3044–3049.

  • Jarvis, P. G., G. B. James, and J. J. Landsberg, 1976: Coniferous Forests. Vegetation and the Atmosphere, J. L. Monteith, Ed., Vol. 2, Academic Press, 171–240.

  • Kustas, W. P., and Coauthors, 1991: An interdisciplinary field study of the energy and water fluxes in the atmosphere–biosphere system over semiarid rangelands: Description and some preliminary results. Bull. Amer. Meteor. Soc.,72, 1683–1705.

  • ——, K. S. Humes, J. M. Norman, and M. S. Moran, 1996: Single- and dual-source modeling of surface energy fluxes with radiometric surface temperature. J. Appl. Meteor.,35, 111–121.

  • ——, X. Zhan, and T. J. Schmugge, 1998: Combining optical and microwave remote sensing for mapping energy fluxes in a semiarid watershed. Remote Sens. Environ.,64, 116–131.

  • Lettau, H. H., and K. Lettau, 1978: Exploring the World’s Driest Climate. IES Rep. 101, University of Wisconsin—Madison, 264 pp. [Available from Schwerdtfeger Library, SSEC, University of Wisconsin—Madison, Madison, WI 53706.].

  • Mahfouf, J. -F., 1991: Analysis of soil moisture from near-surface parameters: A feasibility study. J. Appl. Meteor.,30, 1534–1547.

  • ——, A. O. Manzi, J. Noilhan, H. Giordana, and M. Deque, 1995: The land surface scheme ISBA within the Météo-France climate model ARPEGE. Part I: Implementation and preliminary results. J. Climate,8, 2039–2057.

  • McNaughton, K. J., and T. W. Spriggs, 1986: A mixed-layer model for regional evaporation. Bound.-Layer Meteor.,74, 243–262.

  • Mecikalski, J. R., G. R. Diak, J. M. Norman, and M. C. Anderson, 1997: Evaluation of regional-scale land surface sensible heating using shelter-level measurements of atmospheric temperature combined with analyses of upper-air data. Agric. For. Meteor.,88, 101–110.

  • Monteith, J. L., and M. H. Unsworth, 1990: Principles of Environmental Biophysics. 2d ed. Edward Arnold, 291 pp.

  • Norman, J. M., J.-L. Chen, and N. S. Goel, 1990: Thermal emissivity and infrared temperature dependence of plant canopy architecture and view angle. Proc. Tenth Annual Int. Geoscience Remote Sensing Symp., Vol. 1., IEEE, Piscataway, NJ, 755–756.

  • ——, W. P. Kustas, and K. S. Humes, 1995: A two-source approach for estimating soil and vegetation energy fluxes from observations of directional radiometric surface temperature. Agric. For. Meteor.,77, 263–293.

  • Price, J. C., 1980: The potential of remotely sensed thermal infrared data to infer surface soil moisture and evaporation. Water Resour. Res.,16, 878–895.

  • ——, 1983: Estimating surface temperatures from satellite thermal infrared data. A simple formulation for the atmospheric effect. Remote Sens. Environ.,13, 353–361.

  • Priestley, C. H. B., and R. J. Taylor, 1972: On the assessment of surface heat flux and evaporation using large-scale parameters. Mon. Wea. Rev.,100, 81–92.

  • Redmond, K., 1980: An emissivity parameterization suitable for climate modeling. Mon. Wea. Rev.,108, 663–675.

  • Rosenzweig, C., and F. Abramopoulos, 1997: Land-surface model development for the GISS GCM. J. Climate,10, 2040–2054.

  • Saugier, B., A. Granier, J. Y. Pontailler, E. Dufrene, and D. D. Baldocchi, 1997: Transpiration of a boreal pine forest measured by branch bag, sap flow and micrometeorological methods. Tree Physiol.,17, 511–519.

  • Sellers, P. J., F. G. Hall, G. Asrar, D. E. Strebel, and R. E. Murphy, 1988: The First ISLSCP Field Experiment (FIFE). Bull. Amer. Meteor. Soc.,69, 22–27.

  • ——, S. O. Los, C. J. Tucker, C. O. Justice, D. A. Dazlich, G. J. Collatz, and D. A. Randall, 1996: A revised land surface parameterization (SiB2) for atmospheric GCMs. Part II: The generation of global fields of terrestrial biophysical parameters from satellite data. J. Climate,9, 706–737.

  • Smith, W. L., L. M. Leslie, G. R. Diak, B. M. Goodman, C. S. Velden, G. M. Callan, W. Raymond, and G. S. Wade, 1988: The integration of meteorological satellite imagery and numerical dynamical forecast models. Philos. Trans. Roy. Soc. London,A324, 317–323.

  • Stokes, G. M., and S. E. Schwartz, 1994: The Atmospheric Radiation Measurement (ARM) Program: Programmatic background and design of the cloud and radiation testbed. Bull. Amer. Meteor. Soc.,75, 1201–1221.

  • Suomi, V. E., R. Fox, S. S. Limaye, and W. L. Smith, 1983: McIDAS-III: A modern data access and analysis system. J. Appl. Meteor.,22, 766–778.

  • Tennekes, H., 1973: A model for the dynamics of the inversion above a convective boundary layer. J. Atmos. Sci.,30, 558–567.

  • U.S. Geological Survey, 1995: Conterminous U.S. land cover characteristics dataset: 1990 prototype. CD-ROM, Sioux Falls, SD.

  • Wetzel, P. J., and A. Boone, 1995: A parameterization for land–atmosphere–cloud exchange (PLACE): Documentation and testing of a detailed process model of the partly cloudy boundary layer over heterogeneous land. J. Climate,8, 1810–1837.

  • ——, D. Atlas, and R. Woodward, 1984: Determining soil moisture from geosynchronous satellite infrared data: A feasibility study. J. Climate Appl. Meteor.,23, 375–391.

  • Wu, X., G. R. Diak, C. M. Hayden, and J. A. Young, 1995: Short-range precipitation forecasts using assimilation of simulated satellite water vapor profiles and cloud liquid water. Mon. Wea. Rev.,123, 347–365.

APPENDIX

ALEXI Basic Equation Set

The ALEXI model uses time changes in surface radiometric temperature (δTRAD) to partition the net radiation (Rn) at the earth’s surface into fluxes of sensible (H), latent (LE) and soil (G) heating. The basic equation set comprising the surface component of the ALEXI model is listed below.

Surface radiometric temperature at view angle ϕ:
TRADϕfϕTcfϕTs
Vegetation cover fraction viewed at nadir and at angle ϕ (assuming a spherical leaf angle distribution with leaf area index F):
i1520-0450-38-9-1352-ea2
System, soil, and canopy energy budgets:
i1520-0450-38-9-1352-ea4
Net radiation:
i1520-0450-38-9-1352-ea7
Sensible heat:
i1520-0450-38-9-1352-ea10
Latent heat:
i1520-0450-38-9-1352-ea13
Soil conduction heat:
GRn,s
Here, the subscripts s and c represent soil and canopy flux components, respectively. Temperature subscripts a, ac, c, and s refer to temperatures of the air above the canopy and within the canopy, and the canopy vegetation and soil surface temperatures, respectively.

The form of the ALEXI model described by Anderson et al. (1997) assumed that fluxes from the soil and canopy components of the scene add roughly in parallel to yield the total surface flux. Norman et al. (1995) show that this is a reasonable assumption for sparsely vegetated surfaces. Here, we use a more general resistance formulation, where both the soil and the vegetation influence the microclimate within the canopy air space, as shown in Fig. 1a. The resistances in this network are Ra, the aerodynamic resistance between the canopy and the upper boundary of the model; Rx, the total boundary layer resistance over all leaves in the canopy; and Rs, the resistance through the boundary layer immediately above the soil surface. Mathematical expressions for these resistances are given by Norman et al. (1995).

In these equations, Rn is the net radiation above the canopy, Rn,c is the component absorbed by the canopy, and Rn,s is the component penetrating to the soil surface. Equations (A7)–(A9) are different from those used by Anderson et al. (1997). The simple exponential extinction function used by Anderson et al. (1997) for Rn,s is not appropriate in more sparse canopies where large upward-directed thermal fluxes can originate from the soil. Here, the longwave components of Rn and Rn,s are a function of the thermal radiation from the sky (Ld), the canopy (Lc), and the soil (Ls), and the coefficient of diffuse radiation transmission through the canopy (τc). The shortwave components depend on insolation values above the canopy (Sd) and above the soil surface (Sd,s), and the reflectivity of the soil–canopy system (A) and the soil surface (ρs). Campbell and Norman (1998) provide approximations for τc depending on leaf absorptivity (Table 2) and leaf area index. The symbols Su and Su,s represent the upwelling shortwave radiation above the soil–canopy system and the soil surface, respectively, and Lu and Lu,s are the analogous longwave emissions. Symbol αPT in Eq. (A14) is the Priestly–Taylor coefficient, equal to 1.3. Also, fg is the fraction of green vegetation, Δ is the slope of the curve of saturation vapor pressure versus temperature, and γ is the psychrometric constant (0.066 kPa °C−1).

A simple slab model describing planetary boundary layer dynamics (see Fig. 1b) constrains the upper-boundary condition in air temperature (Ta) used in the surface model component [Eq. (A9)]. McNaughton and Spriggs (1986) give a simplified conservation equation relating the rise in height (z) and potential temperature (θm) of the mixed layer to the time-integrated sensible heating from the surface:
i1520-0450-38-9-1352-ea16
Near the land surface, the mixed layer potential temperature and the air temperature are related by
i1520-0450-38-9-1352-ea17
where p is the atmospheric pressure (kPa) and R/cp = 0.286. The symbols ρ, R, and cp are the density, gas constant, and specific heat capacity, respectively, of air. The early morning PBL potential temperature sounding is denoted as θs.
Using brightness temperature measurements acquired at two times during the morning (t1 and t2), the surface model component [Eqs. (A1)–(A14)] yields instantaneous sensible heat flux estimates H1 and H2. Assuming a linear functional form for H(t), a time-integrated heat flux can be obtained:
i1520-0450-38-9-1352-ea18
The surface and boundary layer components of the model iterate until the sensible heat flux estimates from both components [Eqs. (A16) and (A18)] converge. Anderson et al. (1997) provide further details concerning the solution sequence used in the ALEXI model.

Fig. 1.
Fig. 1.

(a) A schematic description of the surface-layer component of the ALEXI model, and (b) the surface-layer model component is applied at times t1 and t2 during the morning hours, returning instantaneous sensible heat flux estimates H1 and H2. The time-integrated sensible heat flux during this interval serves to heat and grow the PBL. Near-surface air temperatures Ta1 and Ta2 are computed as a boundary condition common to the surface and PBL model components. The line θs represents an early morning potential temperature sounding in the PBL. All symbols in Figs. 1a,b are defined in the appendix.

Citation: Journal of Applied Meteorology 38, 9; 10.1175/1520-0450(1999)038<1352:EFOCSU>2.0.CO;2

Fig. 2.
Fig. 2.

Map of land cover classes used in ALEXI model runs, derived from the prototype 1990 conterminous U.S. land surface characteristics dataset developed by the U.S. Geological Survey (U.S. Geological Survey 1995). Class descriptions are given in Table 2, and are similar to those used in the SiB1 model (Dorman and Sellers 1989).

Citation: Journal of Applied Meteorology 38, 9; 10.1175/1520-0450(1999)038<1352:EFOCSU>2.0.CO;2

Fig. 3.
Fig. 3.

Fractional vegetation cover (1.00 = 100%) input data fields for 12 Jun 1995.

Citation: Journal of Applied Meteorology 38, 9; 10.1175/1520-0450(1999)038<1352:EFOCSU>2.0.CO;2

Fig. 4.
Fig. 4.

Change in surface radiometric temperature (∂T) between 1.5 and 5.5 h past local sunrise for 12 Jun 1995 and 2 Jul 1997 (lhs), and ALEXI model flux estimates of latent heating (W m−2) for 12 Jun 1995 and 2 Jul 1997 (rhs). Blanked (white) regions are too cloudy for ALEXI model evaluation. Left color bar pertains to radiometric temperature changes (∂T), while right color bar scales latent heating.

Citation: Journal of Applied Meteorology 38, 9; 10.1175/1520-0450(1999)038<1352:EFOCSU>2.0.CO;2

Fig. 5.
Fig. 5.

ALEXI model flux estimates of sensible heating (W m−2) for 12 Jun 1995. Stippled regions are too cloudy for ALEXI model evaluation.

Citation: Journal of Applied Meteorology 38, 9; 10.1175/1520-0450(1999)038<1352:EFOCSU>2.0.CO;2

Fig. 6.
Fig. 6.

Fractional vegetation cover (1.00 = 100%) input data fields for 2 Jul 1997.

Citation: Journal of Applied Meteorology 38, 9; 10.1175/1520-0450(1999)038<1352:EFOCSU>2.0.CO;2

Fig. 7.
Fig. 7.

ALEXI model flux estimates of sensible heating (W m−2) for 2 Jul 1997. Stippled regions are too cloudy for ALEXI model evaluation.

Citation: Journal of Applied Meteorology 38, 9; 10.1175/1520-0450(1999)038<1352:EFOCSU>2.0.CO;2

Fig. 8.
Fig. 8.

Comparison between flux estimates from the ALEXI model at 1.5 and 5.5 h past sunrise (gray squares) and measurements made during SGP-97 (lines) at (a) the ARM CART central facility (36.5°N, 97.4°W; measurements from P. Houser); (b) the El Reno site (35.4°N, 98.0°W; measurements from P. Starks, W. Kustas, and J. Prueger); and (c) the Little Washita watershed site (35.0°N, 98.0°W; measurements from T. Meyers). The different symbols correspond to various stations at a given site. (RNET = net radiation at the earth’s surface, G = soil heat flux, H = sensible heat flux, and LE = latent heat flux.)

Citation: Journal of Applied Meteorology 38, 9; 10.1175/1520-0450(1999)038<1352:EFOCSU>2.0.CO;2

Table 1.

Input data fields required by the ALEXI model, and data sources used in this application. Times t1 and t2 are 1.5 and 5.5 h after local sunrise, respectively.

Table 1.
Table 2.

Land cover classification system used by the ALEXI model along with all quantities that vary according to land cover class. This system was developed for the USGS land characteristics database (U. S. Geological Survey 1995); classes 1–13 are the original SiB1 model classes (Dorman and Sellers 1989), and classes 14–20 were added to capture mosaics and wetlands. Also listed are the values of class-defined parameters used by the ALEXI model: maximum and minimum canopy heights (hmin and hmax), leaf absorptivities in the visible and near-infrared wavelengths [αl,vis and αl,nir, used in the Monteith and Unsworth (1990) scheme for hemispherical albedo, A], and characteristic leaf size (s).

Table 2.
Table 3.

Time-integrated (4 h) comparisons between ALEXI model and SGP-97 sites (ground truth) for net radiation (Rnet), sensible heat (H), latent heat (LE) and ground (heat) flux (G). Fluxes represent a 4-h time integral between 1.5 and 5.5 h after local sunrise. Stations are prefixed “CF” for the ARM CART (central facility) site, “ER” for the El Reno site, and “LW” for the Little Washita watershed. Flux units are MJ m−2 h−1 (1 MJ = 106 J).

Table 3.
Save
  • Anderson, M. C., J. M. Norman, G. R. Diak, W. P. Kustas, and J. R. Mecikalski, 1997: A two-source time-integrated model for estimating surface fluxes using thermal infrared remote sensing. Remote Sens. Environ.,60, 195–216.

  • Avissar, R., and T. Schmidt, 1998: An evaluation of the scale at which ground-surface heat flux patchiness affects the convective boundary layer using large-eddy simulations. J. Atmos. Sci.,55, 2666–2689.

  • Becker, F., and Z.-L. Li, 1990: Temperature-independent spectral indices in thermal infrared bands. Remote Sens. Environ.,32, 17–33.

  • Campbell, G. S., and J. M. Norman, 1998: Introduction to Environmental Biophysics. Springer-Verlag, 286 pp.

  • Carlson, T. N., and D. A. J. Ripley, 1997: On the relationship between fractional vegetation cover, leaf area index, and NDVI. Remote Sens. Environ.,62, 241–252.

  • ——, J. K. Dodd, S. G. Benjamin, and J. N. Cooper, 1981: Satellite estimation of the surface energy balance, moisture availability, and thermal inertia. J. Appl. Meteor.,20, 67–87.

  • ——, W. J. Capehart, and R. R. Gillies, 1995: A new look at the simplified method for remote sensing of daily evapotranspiration. Remote Sens. Environ.,54, 161–167.

  • Choudhury, B. J., S. B. Idso, and R. J. Reginato, 1987: Analysis of an empirical model for soil heat flux under a growing wheat crop for estimating evaporation by an infrared-temperature-based energy balance equation. Agric. For. Meteor.,39, 283–297.

  • ——, N. U. Ahmed, S. B. Idso, R. J. Reginato, and C. S. T. Daughtry, 1994: Relations between evaporation coefficients and vegetation indices studied by model simulations. Remote Sens. Environ.,50, 1–17.

  • Cressman, G., 1959: An operational objective analysis system. Mon. Wea. Rev.,87, 367–374.

  • Culf, A. D., 1993: The potential for estimating regional sensible heat flux from convective boundary layer growth. J. Hydrol.,146, 235–244.

  • DeRidder, K., and G. Schayes, 1997: The IAGL land surface model. J. Appl. Meteor.,36, 167–183.

  • Diak, G. R., 1990: Evaluation of heat flux, moisture flux and aerodynamic roughness at the land surface from knowledge of the PBL height and satellite-derived skin temperatures. Agric. For. Meteor.,52, 181–198.

  • ——, and C. Gautier, 1983: Improvements to a simple physical model for estimating insolation from GOES data. J. Climate Appl. Meteor.,22, 505–508.

  • ——, and T. R. Stewart, 1989: Assessment of surface turbulent fluxes using geostationary satellite surface skin temperatures and a mixed layer planetary boundary layer scheme. J. Geophys. Res.,94, 6357–6373.

  • ——, and M. S. Whipple, 1993: Improvements to models and methods for evaluating the land surface energy balance and “effective” roughness using radiosonde reports and satellite-measured“skin” temperatures. Agric. For. Meteor.,63, 189–218.

  • ——, and ——, 1995: A note on estimating surface sensible heat fluxes using surface temperatures measured from a geostationary satellite during FIFE-1989. J. Geophys. Res.,100, 25 453–25 461.

  • ——, D. Kim, M. S. Whipple, and X. Wu, 1992: Preparing for the AMSU. Bull. Amer. Meteor. Soc.,73, 1971–1984.

  • ——, R. M. Rabin, K. P. Gallo, and C. M. Neale, 1995: Regional-scale comparisons of vegetation and soil wetness with surface energy budget properties from satellite and in-situ observations. Remote Sens. Rev.,12, 355–382.

  • ——, W. L. Bland, and J. R. Mecikalski, 1996: A note on first estimates of surface insolation from GOES-8 visible satellite data. Agric. For. Meteor.,82, 219–226.

  • Divakarla, M., 1997: Estimating spatial distributed surface fluxes from satellite data, in-situ measurements and the Cupid model using GIS (energy flux, carbon dioxide flux, biospheric models). Ph.D. dissertation, University of Wisconsin, 199 pp. [Available from University Microfilm, 305 N. Zeeb Rd., Ann Arbor, MI 48106; AAC 9724097.].

  • Dorman, J. L., and P. J. Sellers, 1989: A global climatology of albedo, roughness length and stomatal resistance for atmospheric general circulation models as represented by the Simple Biosphere Model (SiB). J. Appl. Meteor.,28, 833–855.

  • Driedonks, A. G. M., 1982: Models and observations of the growth of the atmospheric boundary layer. Bound.-Layer Meteor.,23, 283–306.

  • Eidenshink, J. C., 1992: The 1990 conterminous U.S. AVHRR data set. Photogr. Eng. Remote Sens.,58, 809–813.

  • Gillies, R. R., J. Cui, T. N. Carlson, W. P. Kustas, and K. S. Humes, 1997: A verification of the “triangle” method for obtaining surface soil water content and energy fluxes from remote measurements of the Normalized Difference Vegetation Index (NDVI) and surface radiant temperature. Int. J. Remote Sens.,18, 3145–3166.

  • Goudriaan, J., 1977: Crop Micrometeorology: A Simulation Study. Center for Agricultural Publishing and Documentation, 249 pp.

  • Hayden, C. M., G. S. Wade, and T. J. Schmit, 1996: Derived product imagery from GOES-8. J. Appl. Meteor.,35, 153–162.

  • ——, T. J. Schmit, and A. J. Schreiner, 1998: The cloud clearing algorithm for GOES product processing. NOAA/NESDIS Tech. Rep., 19 pp. [Available from A. Schreiner, University of Wisconsin—Madison, Madison, WI 53706.].

  • Henderson-Sellers, A., Z.-L. Yang, and R. E. Dickinson, 1993: The Project for Intercomparison of Land-surface Parameterization Schemes. Bull. Amer. Meteor. Soc.,74, 1335–1349.

  • Idso, S. B., T. J. Schmugge, R. D. Jackson, and R. J. Reginato, 1975:The utility of surface temperature measurements for the remote sensing of surface soil water status. J. Geophys. Res.,80, 3044–3049.

  • Jarvis, P. G., G. B. James, and J. J. Landsberg, 1976: Coniferous Forests. Vegetation and the Atmosphere, J. L. Monteith, Ed., Vol. 2, Academic Press, 171–240.

  • Kustas, W. P., and Coauthors, 1991: An interdisciplinary field study of the energy and water fluxes in the atmosphere–biosphere system over semiarid rangelands: Description and some preliminary results. Bull. Amer. Meteor. Soc.,72, 1683–1705.

  • ——, K. S. Humes, J. M. Norman, and M. S. Moran, 1996: Single- and dual-source modeling of surface energy fluxes with radiometric surface temperature. J. Appl. Meteor.,35, 111–121.

  • ——, X. Zhan, and T. J. Schmugge, 1998: Combining optical and microwave remote sensing for mapping energy fluxes in a semiarid watershed. Remote Sens. Environ.,64, 116–131.

  • Lettau, H. H., and K. Lettau, 1978: Exploring the World’s Driest Climate. IES Rep. 101, University of Wisconsin—Madison, 264 pp. [Available from Schwerdtfeger Library, SSEC, University of Wisconsin—Madison, Madison, WI 53706.].

  • Mahfouf, J. -F., 1991: Analysis of soil moisture from near-surface parameters: A feasibility study. J. Appl. Meteor.,30, 1534–1547.

  • ——, A. O. Manzi, J. Noilhan, H. Giordana, and M. Deque, 1995: The land surface scheme ISBA within the Météo-France climate model ARPEGE. Part I: Implementation and preliminary results. J. Climate,8, 2039–2057.

  • McNaughton, K. J., and T. W. Spriggs, 1986: A mixed-layer model for regional evaporation. Bound.-Layer Meteor.,74, 243–262.

  • Mecikalski, J. R., G. R. Diak, J. M. Norman, and M. C. Anderson, 1997: Evaluation of regional-scale land surface sensible heating using shelter-level measurements of atmospheric temperature combined with analyses of upper-air data. Agric. For. Meteor.,88, 101–110.

  • Monteith, J. L., and M. H. Unsworth, 1990: Principles of Environmental Biophysics. 2d ed. Edward Arnold, 291 pp.

  • Norman, J. M., J.-L. Chen, and N. S. Goel, 1990: Thermal emissivity and infrared temperature dependence of plant canopy architecture and view angle. Proc. Tenth Annual Int. Geoscience Remote Sensing Symp., Vol. 1., IEEE, Piscataway, NJ, 755–756.

  • ——, W. P. Kustas, and K. S. Humes, 1995: A two-source approach for estimating soil and vegetation energy fluxes from observations of directional radiometric surface temperature. Agric. For. Meteor.,77, 263–293.

  • Price, J. C., 1980: The potential of remotely sensed thermal infrared data to infer surface soil moisture and evaporation. Water Resour. Res.,16, 878–895.

  • ——, 1983: Estimating surface temperatures from satellite thermal infrared data. A simple formulation for the atmospheric effect. Remote Sens. Environ.,13, 353–361.

  • Priestley, C. H. B., and R. J. Taylor, 1972: On the assessment of surface heat flux and evaporation using large-scale parameters. Mon. Wea. Rev.,100, 81–92.

  • Redmond, K., 1980: An emissivity parameterization suitable for climate modeling. Mon. Wea. Rev.,108, 663–675.

  • Rosenzweig, C., and F. Abramopoulos, 1997: Land-surface model development for the GISS GCM. J. Climate,10, 2040–2054.

  • Saugier, B., A. Granier, J. Y. Pontailler, E. Dufrene, and D. D. Baldocchi, 1997: Transpiration of a boreal pine forest measured by branch bag, sap flow and micrometeorological methods. Tree Physiol.,17, 511–519.

  • Sellers, P. J., F. G. Hall, G. Asrar, D. E. Strebel, and R. E. Murphy, 1988: The First ISLSCP Field Experiment (FIFE). Bull. Amer. Meteor. Soc.,69, 22–27.

  • ——, S. O. Los, C. J. Tucker, C. O. Justice, D. A. Dazlich, G. J. Collatz, and D. A. Randall, 1996: A revised land surface parameterization (SiB2) for atmospheric GCMs. Part II: The generation of global fields of terrestrial biophysical parameters from satellite data. J. Climate,9, 706–737.

  • Smith, W. L., L. M. Leslie, G. R. Diak, B. M. Goodman, C. S. Velden, G. M. Callan, W. Raymond, and G. S. Wade, 1988: The integration of meteorological satellite imagery and numerical dynamical forecast models. Philos. Trans. Roy. Soc. London,A324, 317–323.

  • Stokes, G. M., and S. E. Schwartz, 1994: The Atmospheric Radiation Measurement (ARM) Program: Programmatic background and design of the cloud and radiation testbed. Bull. Amer. Meteor. Soc.,75, 1201–1221.

  • Suomi, V. E., R. Fox, S. S. Limaye, and W. L. Smith, 1983: McIDAS-III: A modern data access and analysis system. J. Appl. Meteor.,22, 766–778.

  • Tennekes, H., 1973: A model for the dynamics of the inversion above a convective boundary layer. J. Atmos. Sci.,30, 558–567.

  • U.S. Geological Survey, 1995: Conterminous U.S. land cover characteristics dataset: 1990 prototype. CD-ROM, Sioux Falls, SD.

  • Wetzel, P. J., and A. Boone, 1995: A parameterization for land–atmosphere–cloud exchange (PLACE): Documentation and testing of a detailed process model of the partly cloudy boundary layer over heterogeneous land. J. Climate,8, 1810–1837.

  • ——, D. Atlas, and R. Woodward, 1984: Determining soil moisture from geosynchronous satellite infrared data: A feasibility study. J. Climate Appl. Meteor.,23, 375–391.

  • Wu, X., G. R. Diak, C. M. Hayden, and J. A. Young, 1995: Short-range precipitation forecasts using assimilation of simulated satellite water vapor profiles and cloud liquid water. Mon. Wea. Rev.,123, 347–365.

  • Fig. 1.

    (a) A schematic description of the surface-layer component of the ALEXI model, and (b) the surface-layer model component is applied at times t1 and t2 during the morning hours, returning instantaneous sensible heat flux estimates H1 and H2. The time-integrated sensible heat flux during this interval serves to heat and grow the PBL. Near-surface air temperatures Ta1 and Ta2 are computed as a boundary condition common to the surface and PBL model components. The line θs represents an early morning potential temperature sounding in the PBL. All symbols in Figs. 1a,b are defined in the appendix.

  • Fig. 2.

    Map of land cover classes used in ALEXI model runs, derived from the prototype 1990 conterminous U.S. land surface characteristics dataset developed by the U.S. Geological Survey (U.S. Geological Survey 1995). Class descriptions are given in Table 2, and are similar to those used in the SiB1 model (Dorman and Sellers 1989).

  • Fig. 3.

    Fractional vegetation cover (1.00 = 100%) input data fields for 12 Jun 1995.

  • Fig. 4.

    Change in surface radiometric temperature (∂T) between 1.5 and 5.5 h past local sunrise for 12 Jun 1995 and 2 Jul 1997 (lhs), and ALEXI model flux estimates of latent heating (W m−2) for 12 Jun 1995 and 2 Jul 1997 (rhs). Blanked (white) regions are too cloudy for ALEXI model evaluation. Left color bar pertains to radiometric temperature changes (∂T), while right color bar scales latent heating.

  • Fig. 5.

    ALEXI model flux estimates of sensible heating (W m−2) for 12 Jun 1995. Stippled regions are too cloudy for ALEXI model evaluation.

  • Fig. 6.

    Fractional vegetation cover (1.00 = 100%) input data fields for 2 Jul 1997.

  • Fig. 7.

    ALEXI model flux estimates of sensible heating (W m−2) for 2 Jul 1997. Stippled regions are too cloudy for ALEXI model evaluation.

  • Fig. 8.

    Comparison between flux estimates from the ALEXI model at 1.5 and 5.5 h past sunrise (gray squares) and measurements made during SGP-97 (lines) at (a) the ARM CART central facility (36.5°N, 97.4°W; measurements from P. Houser); (b) the El Reno site (35.4°N, 98.0°W; measurements from P. Starks, W. Kustas, and J. Prueger); and (c) the Little Washita watershed site (35.0°N, 98.0°W; measurements from T. Meyers). The different symbols correspond to various stations at a given site. (RNET = net radiation at the earth’s surface, G = soil heat flux, H = sensible heat flux, and LE = latent heat flux.)

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 568 116 8
PDF Downloads 231 60 7