1. Introduction
The downward components of Rn are controlled by solar zenith angle (i.e., time of day, season, and latitude), cloud amount, atmospheric water vapor amount, and aerosol loading; in turn, Rn demonstrates a substantial daily and seasonal variation. The upward components of Rn are controlled by ground surface characteristics including snow/ice coverage, vegetation coverage, and soil moisture content (Wang et al. 2005b, 2007b). Most previous studies only focus on the changes in shortwave radiation (e.g., Pinker et al. 2005; Wild et al. 2005), presumably because longwave radiation is not conventionally measured. The objective of this study is to develop a new method to estimate Rn, especially for climate study.
Numerous studies estimated L↓ using conventional meteorological observations (Malek 1997; Niemelä et al. 2001; Jin et al. 2006; Bilbao and de Miguel 2007; Kjaersgaard et al. 2007b; Lhomme et al. 2007). Results to date show that most methods require long-term longwave radiation observations for local calibration (Jin et al. 2006; Bilbao and de Miguel 2007; Kjaersgaard et al. 2007a; Lhomme et al. 2007). Air temperature and water vapor profile retrievals from satellite observations have also been used to estimate L↓ (Ellingson 1995; Diak et al. 2004; Bisht et al. 2005; Zhou et al. 2007). Wang and Liang (2008) estimated L↓ using only Moderate Resolution Imaging Spectroradiometer (MODIS) top of atmosphere radiance, but their method is suitable only for clear-sky conditions. Quantifying cloud effects is the main difficulty in estimating L↓ under cloudy conditions (Ellingson 1995). Satellite sensors can monitor the temperature of the cloud top; however, the cloud-base temperature is the critical parameter controlling L↓ at the earth’s surface under cloudy conditions. Accuracies of the various methods vary, with the error ranging from about 10 to 20 W m−2 for instantaneous clear-sky L↓, degrading to 20–40 W m−2 for cloudy conditions (Diak et al. 2004).
Liang (2004), Wang et al. (2005a), Zhang et al. (2007), and Wang and Liang (2008) calculated L↑ from surface temperature Ts and broadband emissivity using the Stefan–Boltzmann law. Wang et al. (2005a) and Jin and Liang (2006) developed methods to estimate broadband emissivity and L↑ from MODIS Ts products. However, these methods are unsuitable under cloudy conditions when Ts is unavailable.
Rather than estimating L↓ and L↑ individually to estimate Rn, we propose to estimate Rn directly. Long-term global surface solar shortwave radiation measurements enable us to develop a new method to estimate Rn. Kjaersgaard et al. (2007a) compared six existing models for calculating daytime Rn from solar shortwave radiation using meteorological data at two temperate sites. They concluded that local calibration of the models with at least 5 years of data is essential to obtain stable calibration coefficients. However, all six methods failed to consider surface characteristics, such as vegetation cover fraction, that have substantial effects on surface energy partitioning into latent heat and sensible heat fluxes (Wang et al. 2007c; Wang and Liang 2008), which, in turn, affect Ts and L↑ (Wang et al. 2006, 2007b).
Therefore, based on analysis of available long-term measurements, we propose a robust method to estimate Rn from solar radiation data using satellite and conventional meteorological observations. Surface incident solar radiation is conventionally observed globally (e.g., Gilgen and Ohmura 1999; Wild et al. 2005) and can also be estimated from satellite observations (e.g., Diak and Gautier 1983; Pinker and Ewing 1985; Dedieu et al. 1987; Li et al. 1993; Ceballos et al. 2004; Liang et al. 2006, 2007). The strength of this method is that it accurately estimates Rn for a wide variety of land cover types and climates and a range of surface elevations, without local calibration.
2. Data
The two most important criteria for data and site selection in this study are 1) the quality of Rn data and the corresponding meteorological observations and 2) the long-term continuity of data; surface elevation, land cover types, and climate zone of the sites are also considered. The 24 sites selected are the 14 Energy Balance Bowen Ratio sites of the Enhanced Facility of the Atmospheric Radiation Measurement (ARM) Program supported by the U.S. Department of Energy, 7 Surface Radiation Budget Network (SURFRAD) sites supported by the National Oceanic and Atmospheric Administration (NOAA), 1 AmeriFlux forest site, 1 Asian Automatic Weather Station Network Project (ANN) site supported by the Global Energy and Water Cycle Experiment (GEWEX) Asian Monsoon Experiment (GAME), and 1 site supported by the Japanese Experiment on Asian Monsoons (JEXAM) and the Frontier Observational Research System for Global Change (FORSGC), University of Tokyo, and Kyoto University (Table 1).
The nature of the ARM, SURFRAD, AmeriFlux, ANN, and JEXAM projects ensures that the data are some of the best quality currently available. The quality of the data used in this study is carefully checked, and quality control flags are supplied before the data are released by data centers. The primary objective of SURFRAD is to support climate research with accurate, continuous, long-term measurements of the surface radiation budget over the United States (http://www.srrb.noaa.gov/), whereas ARM uses clustered measurements over a limited area for process-oriented studies (http://www.arm.gov/). GAME ANN was implemented to understand the role of the Asian monsoon in the global energy and water cycle (http://aan.suiri.tsukuba.ac.jp/aan.html). AmeriFlux aims to quantify spatial and temporal variation in carbon storage in plants and soils, and exchanges of carbon, water, and energy in major vegetation types; the accuracy of radiation measurements is not as high as that of SURFRAD measurements.
The data used in this study are highly accurate in large part because of well-calibrated sensors. Measurement accuracy is about 6% for shortwave radiation measurements and about 2.5% for longwave radiation for observations at the sites. The SURFRAD project and Gaize sites (http://www.srrb.noaa.gov/surfrad/getcals.html; Wang et al. 2004, 2005b) carefully calibrate their sensors annually. All SURFRAD and ARM data are carefully checked, and quality-control information is supplied with the data (ftp://ftp.srrb.noaa.gov/pub/data/surfrad/ and http://www.archive.arm.gov/). Internet sites listed in this paragraph provide sensor information and measurement accuracies for all sites except the Gaize site; Wang et al. (2004, 2005b) provide information on the Gaize site.
We have successfully used data from SURFRAD, ARM, AmeriFlux, and the Gaize site to validate radiation and albedo retrievals and to estimate physical processes confirming the data quality and accuracy. We have successfully used radiation and meteorological data collected by ARM and AmeriFlux to estimate evapotranspiration (Wang et al. 2007c; Wang and Liang 2008) and evaporative fraction and light use efficiency (Wang et al. 2006, 2008). Wang et al. (2008) used SURFRAD radiation measurements to successfully validate satellite shortwave and longwave radiation retrievals (Wang and Liang 2008, manuscript submitted to Remote Sens. Environ.). Wang et al. (2004) successfully used Gaize-site shortwave radiation measurements to validate satellite albedo retrievals, and Wang et al. (2005a) used Gaize longwave radiation measurements to develop a satellite L↑ algorithm.
We selected sites for this study that include a number of land cover types, climates, and elevations ranging from 98 to 4700 m to ensure the applicability of our method (Table 1). All sites are located in the United States except for two sites located on the Tibetan Plateau, China. Land cover types include desert, semidesert, cropland, grassland, and forests. Most sites selected for this study are pasture and grassland. Collection of accurate S↑ and L↑ over relatively low stand pasture and grassland vegetation is considerably easier than over high-stand forested areas, and therefore only one forest site is included in this study. Because high-quality desert datasets are scarce because of the difficulty in collecting long-term continuous measurements over desert, only one desert site (Desert Rock, Nevada) and one semidesert (Gaize, China) site are included in this study.
The problem with deserts is that sensors are not consistently maintained for long time periods. Therefore, long-term continuous desert measurements are seldom available. Previous experience shows that the strong daytime solar radiation and low nighttime temperatures affect the performance of pyrgeometers (Wang et al. 2007a). To measure upward shortwave radiation and longwave radiation accurately, the measurement height is required to be about 10 m above the surface. For forest sites (typically 20–40 m above the ground surface), the tower is required to be very high. Therefore, most forest sites do not supply upward radiation measurements.
3. Methods
Unlike longwave radiation, surface incident shortwave solar radiation S↓ is conventionally observed globally. The Global Energy Budget Archive (GEBA) project collects long-term solar shortwave radiation at more than 1500 stations worldwide (Gilgen and Ohmura 1999; Wild et al. 2005). GEBA also evaluates data quality. The ratio of Rn to daytime net shortwave radiation (Rn/Sn) is used to estimate daytime Rn from shortwave radiation measurements. Daytime is any time during which Sn exceeds 20 W m−2. GEBA S↓ measurements enable calculation of Sn with the help of surface albedo. At this time, long-term albedo is available globally at a high spatial resolution of several kilometers and a relatively high temporal resolution of one-half of a month (e.g., Pinty et al. 2000; Pokrovsky et al. 2003; Schaaf et al. 2002; Strugnell and Lucht 2001). Figure 1 shows an example of a time series of Rn/Sn that has a substantial seasonal variation. We selected albedo from ground measurements to keep the scale of ground-measured and predicted Rn consistent. Therefore, we can get Rn by estimating the ratio Rn/Sn.
In our previous studies, we showed that vegetation cover fraction, which can be quantified by NDVI or EVI (Tucker et al. 1985), substantially affects surface energy partitioning into latent heat and sensible heat fluxes (Wang et al. 2007c; Wang and Liang 2008), which in turn affects Ts and L↑ (Wang et al. 2006, 2007b). Global long-term NDVI at high spatial resolution (several kilometers) and a relatively high temporal resolution (week or one-half of a month) is available from multiple satellite sensors and data centers (e.g., Los et al. 2005; Tucker et al. 2005; Jiang et al. 2008; Swinnen and Veroustraete 2008).
To parameterize the ratio Rn/Sn and to estimate Rn, we calculate the correlation coefficients between the ratio and daily average, daily maximum, and daily minimum air temperature (Ta or Ts), daily Ta (or Ts) range, relative humidity (RH), and NDVI and EVI. Table 2 lists the correlation coefficients. Among the various measures, daily minimum temperature has the highest correlation coefficient with the ratio, followed by RH, vegetation index, and daily Ts (or Ta) range. Relative humidity, air temperature, and diurnal temperature (Ta or Ts) range (closely related to cloud cover) are the key parameters controlling L↓. The Ts, diurnal temperature range (Ta or Ts), and vegetation indices are the key parameters controlling surface energy partitioning and, in turn, L↑ (Wang et al. 2006, 2007c; Wang and Liang 2008). Diurnal temperature (Ts or Ta) range (DTaR or DTsR), the difference between daily maximum temperature and daily minimum temperature, accounts for the complex effects of clouds, surface daytime temperature dynamics, and soil moisture on Rn.
The relationships between Rn/Sn and DTaR, Ta min, VI, and RH are not exactly linear (Fig. 2). Thus, none of the parameters of DTaR, Ta min, VI, and RH can individually account for the variance in Rn/Sn, although the four parameters in combination estimate the variance in Rn/Sn with greater accuracy. The results described in section 4 substantiate this. We also performed the regression analysis with the square of the terms in addition to the linear terms without substantially improving the results. Therefore, only linear regressions are used in this paper.
4. Results analysis
The data collected at 24 sites have two purposes. We used the Amdo, Bondville, Desert Rock, EF02, EF07, EF12, EF15, EF18, EF19, Fort Peck, “Penn State,” and Sioux Falls data to derive the coefficients in Eq. (6), and we validated the coefficients with the Boulder, Gaize, EF04, EF08, EF09, EF13, EF20, EF22, EF26, EF27, Goodwin Creek, and Wind River data. Table 3 lists the derived coefficients.
As expected, the coefficients listed in Table 3 indicate the importance of NDVI (or EVI) in accounting for the variation in L↑ and Rn. Air humidity and temperature (Ta or Ts) are generally considered to account for most of the variation in L↓. This is also the case in this study, because RH and Ta (Ts) are the dominant terms for Rn in Eq. (6) (Table 3). DTR is a measure of the effect of soil moisture on Ts dynamics and L↑. Therefore, a2 is more important when Ts is used in Eq. (6). In statistical terms, Ta (Ts), DTaR (DTsR), and RH are all related to cloud cover and its effects on Rn. The focus of this study is on estimating daytime average Rn under both clear- and cloudy-sky conditions because variations in percentage cloud cover are likely to occur over the course of a day. Figure 3 shows an example of measured and predicted Rn from daily minimum Ts, DTsR, NDVI, and RH. Equation (6) predicts Rn using the same coefficients at the 24 sites. Table 4 summarizes the statistical parameters of the comparison of predicted and measured Rn at the 24 sites.
Equation (6) and the Table 3 coefficients accurately estimate Rn for a range of land cover types, surface elevations, and climates without local calibration (Table 4). The bias varies from −7.8 to +9.7 W m−2 (±3% in relative value) for different sites. The root-mean-square error (RMSE) varies from 12.8 to 21 W m−2 (from 5% to 9% in relative value) for different sites and an average of 16.9 W m−2 (6% in relative value) for all sites, and the correlation coefficient is about 0.99 for all sites.
Wang et al. (2007c) and Wang and Liang (2008) demonstrated that EVI better quantifies vegetation cover fraction because EVI is less dependent on soil type. Therefore, EVI may be more accurate in parameterizing Rn; however, EVI is only available after 2000 (Salomonson et al. 1989). NDVI is used to estimate Rn when EVI is unavailable. Our previous studies also show that Ts is directly related to L↑, whereas the relationship between Ta and L↑ is indirect (Wang et al. 2005b, 2007b). However, satellite Ts retrieval is not available under cloudy conditions. Therefore, we also provide an equation using Ta.. Table 3 also shows that other combinations of temperatures, vegetation index, and relative humidity produce similar results. The results with Ts demonstrate slightly better overall statistical parameters.
Changes in Rn impact a host of factors, including temperature, precipitation, meteorological patterns, and sea level (Charlson et al. 2005). Therefore, predicting long-term variation in Rn, such as year-to-year variation and decadal variation, is important. To examine the applicability of the method to climate study, we compare the annual abnormality (year average subtracted from multiyear average) in measured and predicted daytime Rn. The results, shown in Fig. 4, demonstrate that Eq. (6) predicts the annual abnormality in daytime Rn accurately, with a correlation coefficient between measured and predicted annual abnormality as high as 0.91. This suggests that Eq. (6) accurately monitors long-term change in Rn.
5. Conclusions and discussion
Changes in Rn broadly affect the earth’s climates, the hydrological cycle, and plant photosynthesis. Existing studies focus on solar shortwave radiation (to the exclusion of longwave radiation), because it is readily available from conventional measurements (whereas longwave radiation is not). Methods to estimate longwave radiation and Rn are essential for climate studies. Current methods to estimate L↓ suffer from the difficulty of quantifying cloud effects. The estimation of L↑ requires Ts as a key input datum, which is unavailable under cloudy conditions. In addition, previous studies demonstrate that it is essential to locally calibrate the existing models with at least 5 years of data to achieve stable calibration coefficients.
Rather than estimating L↓ and L↑ separately, we estimate Rn directly. The new method accurately estimates daytime Rn from solar radiation measurements using data collected at 24 sites over a range of land cover types, climate zones, and surface elevations. The method is based on our previous research on land energy partitioning into surface latent heat flux and sensible heat flux and their interactions with Ts and L↑ (Wang et al. 2006, 2007b,c; Wang and Liang 2008). We use NDVI (or EVI), which are parameters related to vegetation cover fraction and soil moisture, to estimate Rn from solar radiation measurements. The results show that an equation to estimate Rn is suitable for all sites. The bias varies from −7.8 to +9.7 W m−2 (±3% in relative value), and RMSE varies from 12.8 to 21 W m−2 (from 5% to 9% in relative value) with an average of 16.9 W m−2 (6% in relative value) for all sites, with a correlation coefficient of about 0.99 for all sites. The accuracy improves upon those in previous studies (e.g., Diak et al. 2004; Kjaersgaard et al. 2007a). The proposed method is suitable for a range of land cover types, surface elevations, and climate zones without local calibration.
Another advantage of the proposed method is that it relies only on conventional meteorological observations and global available satellite data. DTR is used in the parameterization proposed here. When Ts is used in the proposed method, the effect of DTsR is very small, so that it is not necessary to incorporate it in the parameterization. When using air temperature, DTaR accounts for about 3% variation of Rn, by our calculations.
The method is based on the analysis of daily net radiation, daily meteorological observations, and satellite daily Ts or 16-day vegetation indices. Our validation results demonstrate that this method works well for predicting daytime daily Rn and annual Rn variation. The correlation coefficient between the measured and predicted annual abnormality is as high as 0.91, indicating that this method accurately estimates long-term variation in Rn. At this time, the long-term datasets of the input data required by the proposed method are globally available. These data can be used to derive global, long-term Rn. For example, GEBA provides S↑ at more than 1500 stations (Gilgen and Ohmura 1999; Wild et al. 2005). At this time, long-term albedo is available globally at high spatial resolution (several kilometers) and a relatively high temporal resolution (one-half of a month) (e.g., Pinty et al. 2000; Pokrovsky et al. 2003; Schaaf et al. 2002; Strugnell and Lucht 2001). A number of satellite sensors and data centers provide global long-term NDVI datasets at high spatial resolution (several kilometers) and a relatively high temporal resolution (week or one-half of a month) (e.g., Los et al. 2005; Tucker et al. 2005; Jiang et al. 2008; Swinnen and Veroustraete 2008).
To validate the proposed data, we used the site measurements and satellite observations at the spatial resolution of 1 km. We believe our method will work well for relatively large spatial resolutions because the relationships between Rn and Sn, water vapor, vegetation coverage, and temperature do not significantly vary with spatial scales.
Acknowledgments
Shortwave radiation, net radiation, and corresponding meteorological observations were obtained from the NOAA SURFRAD project (ftp://ftp.srrb.noaa.gov/pub/data/surfrad/), the ARM Program of the U.S. Department of Energy (http://www.archive.arm.gov/), the AmeriFlux network (http://public.ornl.gov/ameriflux/data-get.cfm), and the GAME ANN project (http://aan.suiri.tsukuba.ac.jp/aan-center.html). MODIS satellite data were obtained online (http://redhook.gsfc.nasa.gov/~imswww/pub/imswelcome/plain.html). This study was supported in part by NASA Grant NNX08DC53G and NOAA Grant NA07NES4400001.
REFERENCES
Bilbao, J., and A. H. D. E. Miguel, 2007: Estimation of daylight downward longwave atmospheric irradiance under clear-sky and all-sky conditions. J. Appl. Meteor. Climatol., 46 , 878–889.
Bisht, G., V. Venturini, S. Islama, and L. Jiang, 2005: Estimation of the net radiation using MODIS (Moderate Resolution Imaging Spectroradiometer) data for clear sky days. Remote Sens. Environ., 97 , 52–67.
Ceballos, J. C., M. J. Bottino, and J. M. de Souza, 2004: A simplified physical model for assessing solar radiation over Brazil using GOES 8 visible imagery. J. Geophys. Res., 109 , D02211. doi:10.1029/2003JD003531.
Charlson, R. J., F. P. J. Valero, and J. H. Seinfeld, 2005: In search of balance. Science, 308 , 806–807.
Dedieu, G., P. Y. Deschamps, and Y. H. Kerr, 1987: Satellite estimation of solar irradiance at the surface of the earth and of surface albedo using a physical model applied to Meteosat data. J. Climate Appl. Meteor., 26 , 79–87.
Diak, G. R., and C. Gautier, 1983: Improvements to a simplified physical model for estimating insolation from GOES data. J. Climate Appl. Meteor., 22 , 505–508.
Diak, G. R., J. R. Mecikalski, M. C. Anderson, J. M. Norman, W. P. Kustas, R. D. Torn, and R. L. DeWolf, 2004: Estimating land surface energy budgets from space: Review and current efforts at the University of Wisconsin—Madison and USDA–ARS. Bull. Amer. Meteor. Soc., 85 , 65–78.
Ellingson, R. G., 1995: Surface longwave fluxes from satellite observations: A critical review. Remote Sens. Environ., 51 , 89–97.
Gilgen, H., and A. Ohmura, 1999: The Global Energy Balance Archive. Bull. Amer. Meteor. Soc., 80 , 831–850.
Huete, A., K. Didan, T. Miura, E. P. Rodriguez, X. Gao, and L. G. Ferreira, 2002: Overview of the radiometric and biophysical performance of the MODIS vegetation indices. Remote Sens. Environ., 83 , 195–213.
Jiang, L., J. D. Tarpley, K. E. Mitchell, S. Zhou, F. N. Kogan, and W. Guo, 2008: Adjusting for long-term anomalous trends in NOAA’s global vegetation index data sets. IEEE Trans. Geosci. Remote Sens., 46 , 409–422.
Jin, M., and S. Liang, 2006: Improved emissivity parameterization for land surface modeling using global remote sensing observations. J. Climate, 19 , 2867–2881.
Jin, X., D. Barber, and T. Papakyriakou, 2006: A new clear-sky downward longwave radiative flux parameterization for Arctic areas based on rawinsonde data. J. Geophys. Res., 111 , D24104. doi:10.1029/2005JD007039.
Kjaersgaard, J. H., R. H. Cuenca, F. L. Plauborg, and S. Hansen, 2007a: Long-term comparisons of net radiation calculation schemes. Bound.-Layer Meteor., 123 , 417–431.
Kjaersgaard, J. H., F. L. Plauborg, and S. Hansen, 2007b: Comparison of models for calculating daytime long-wave irradiance using long term data set. Agric. For. Meteor., 143 , 49–63.
Lhomme, J. P., J. J. Vacher, and A. Rocheteau, 2007: Estimating downward long-wave radiation on the Andean Altiplano. Agric. For. Meteor., 145 , 139–148.
Li, Z., H. G. Leighton, K. Masuda, and T. Takashima, 1993: Estimation of SW flux absorbed at the surface from TOA reflected flux. J. Climate, 6 , 317–330.
Liang, S., 2004: Quantitative Remote Sensing of Land Surfaces. John Wiley and Sons, 534 pp.
Liang, S., T. Zheng, R. Liu, H. Fang, S. C. Tsay, and S. Running, 2006: Estimation of incident photosynthetically active radiation from MODIS data. J. Geophys. Res., 111 , D15208. doi:10.1029/2005JD006730.
Liang, S., T. Zheng, D. Wang, K. Wang, R. Liu, S. C. Tsay, S. Running, and J. Townshend, 2007: Mapping high-resolution incident photosynthetically active radiation over land from polar-orbiting and geostationary satellite data. Photogramm. Eng. Remote Sens., 73 , 1085–1089.
Los, S. O., P. R. J. North, W. M. F. Grey, and M. J. Barnsley, 2005: A method to convert AVHRR normalized difference vegetation index time series to a standard viewing and illumination geometry. Remote Sens. Environ., 99 , 400–411.
Malek, E., 1997: Evaluation of effective atmospheric emissivity and parameterization of cloud at local scale. Atmos. Res., 45 , 41–54.
Niemelä, S., P. Räisänen, and H. Savijärvi, 2001: Comparison of surface radiative flux parameterizations: I. Longwave radiation. Atmos. Res., 58 , 1–18.
Pinker, R. T., and J. A. Ewing, 1985: Modeling surface solar radiation: Model formulation and validation. J. Climate Appl. Meteor., 24 , 389–401.
Pinker, R. T., B. Zhang, and E. G. Dutton, 2005: Do satellites detect trends in surface solar radiation? Science, 308 , 850–854.
Pinty, B., and Coauthors, 2000: Surface albedo retrieval from Meteosat 1. Theory. J. Geophys. Res., 105 , 18099–18112.
Pokrovsky, I., O. Pokrovsky, and J-L. Roujean, 2003: Development of an operational procedure to estimate surface albedo from the SEVIRI/MSG observing system by using POLDER BRDF measurements I. Data quality control and accumulation of information corresponding to the IGBP land cover classes. Remote Sens. Environ., 87 , 198–214.
Salomonson, V., W. Barnes, P. Maymon, H. Montgomery, and H. Ostrow, 1989: MODIS: advanced facility instrument for studies of the earth as a system. IEEE Trans. Geosci. Remote Sens., 27 , 145–153.
Schaaf, C. B., and Coauthors, 2002: First operational BRDF, albedo and nadir reflectance products from MODIS. Remote Sens. Environ., 83 , 135–148.
Strugnell, N. C., and W. Lucht, 2001: An algorithm to infer continental-scale albedo from AVHRR data, land cover class, and field observations of typical BRDFs. J. Climate, 14 , 1360–1376.
Swinnen, E., and F. Veroustraete, 2008: Extending the SPOT-VEGETATION NDVI time series (1998–2006) back in time with NOAA-AVHRR data (1985–1998) for southern Africa. IEEE Trans. Geosci. Remote Sens., 46 , 558–572.
Tucker, C. J., J. R. G. Townshend, and T. E. Goff, 1985: African land-cover classification using satellite data. Science, 227 , 369–375.
Tucker, C. J., and Coauthors, 2005: An extended AVHRR 8-km dataset compatible with MODIS and SPOT vegetation NDVI data. Int. J. Remote Sens., 26 , 4485–4498.
Van Leeuwen, W., A. Huete, and T. Laing, 1999: MODIS vegetation index compositing approach: A prototype with AVHRR data. Remote Sens. Environ., 69 , 264–280.
Wang, K., and S. Liang, 2008: An improved method for estimating global evapotranspiration based on satellite determination of surface net radiation, vegetation index, temperature, and soil moisture. J. Hydrometeor., 9 , 712–727.
Wang, K., J. Liu, X. Zhou, M. Sparrow, M. Ma, Z. Sun, and W. Jiang, 2004: Validation of the MODIS global land surface albedo product using ground measurements in a semidesert region on the Tibetan Plateau. J. Geophys. Res., 109 , D05107. doi:10.1029/2003JD004229.
Wang, K., Z. Wan, P. Wang, M. Sparrow, J. Liu, X. Zhou, and S. Haginoya, 2005a: Estimation of surface long wave radiation and broadband emissivity using Moderate Resolution Imaging Spectroradiometer (MODIS) land surface temperature/emissivity products. J. Geophys. Res., 110 , D11109. doi:10.1029/2004JD005566.
Wang, K., P. Wang, J. Liu, S. Michael, H. Shigenori, and X. Zhou, 2005b: Variation of surface albedo and soil thermal parameters with soil moisture content at a semi-desert site on the western Tibetan Plateau. Bound.-Layer Meteor., 116 , 117–129.
Wang, K., Z. Li, and M. Cribb, 2006: Estimation of evaporative fraction from a combination of day and night land surface temperature and NDVI: A new method to determine the Priestley–Taylor parameter. Remote Sens. Environ., 102 , 293–305.
Wang, K., Z. Wan, P. Wang, M. Sparrow, J. Liu, and S. Haginoya, 2007a: Evaluation and improvement of the MODIS land surface temperature/emissivity products using ground-based measurements at a semi-desert site on the western Tibetan Plateau. Int. J. Remote Sens., 28 , 2549–2565.
Wang, K., J. Wang, P. Wang, M. Sparrow, J. Yang, and H. Chen, 2007b: Influences of urbanization on surface characteristics as derived from the Moderate-Resolution Imaging Spectroradiometer: A case study for the Beijing metropolitan area. J. Geophys. Res., 112 , D22S06. doi:10.1029/2006JD007997.
Wang, K., P. Wang, Z. Li, M. Cribb, and M. Sparrow, 2007c: A simple method to estimate actual evapotranspiration from a combination of net radiation, vegetation index, and temperature. J. Geophys. Res., 112 , D15107. doi:10.1029/2006JD008351.
Wang, K., R. E. Dickinson, and S. Liang, 2008: Observational evidence on the effects of clouds and aerosols on net ecosystem exchange and evapotranspiration. Geophys. Res. Lett., 35 , L10401. doi:10.1029/2008GL034167.
Wang, W., and S. Liang, 2008: Estimating high spatial resolution clear-sky surface downwelling longwave radiation and net longwave radiation from MODIS Data. Remote Sens. Environ., in press.
Wang, W., S. Liang, and J. A. Augustine, 2008: Estimating clear-sky land surface upwelling longwave radiation from MODIS data. IEEE Trans. Geosci. Remote Sens., in press.
Wild, M., and Coauthors, 2005: From dimming to brightening: Decadal changes in solar radiation at Earth’s surface. Science, 308 , 847–850.
Zhang, Y., W. B. Rossow, and P. W. Stackhouse Jr., 2007: Comparison of different global information sources used in surface radiative flux calculation: Radiative properties of the surface. J. Geophys. Res., 112 , D01102. doi:10.1029/2005JD007008.
Zhou, Y., D. P. Kratz, A. C. Wilber, S. K. Gupta, and R. D. Cess, 2007: An improved algorithm for retrieving surface downwelling longwave radiation from satellite measurements. J. Geophys. Res., 112 , D15102. doi:10.1029/2006JD008159.
An example of the time series of the ratio Rn/Sn of surface daytime Rn to net shortwave radiation Sn using data collected at Hillsboro, KS (EF02).
Citation: Journal of Applied Meteorology and Climatology 48, 3; 10.1175/2008JAMC1959.1
An example of the time series of the ratio Rn/Sn of surface daytime Rn to net shortwave radiation Sn using data collected at Hillsboro, KS (EF02).
Citation: Journal of Applied Meteorology and Climatology 48, 3; 10.1175/2008JAMC1959.1
An example of the time series of the ratio Rn/Sn of surface daytime Rn to net shortwave radiation Sn using data collected at Hillsboro, KS (EF02).
Citation: Journal of Applied Meteorology and Climatology 48, 3; 10.1175/2008JAMC1959.1
An example of the ratio Rn/Sn as a function of DTaR, (Ta min), NDVI, and RH using data collected at Earlsboro, OK (EF27) from 2003 to 2006.
Citation: Journal of Applied Meteorology and Climatology 48, 3; 10.1175/2008JAMC1959.1
An example of the ratio Rn/Sn as a function of DTaR, (Ta min), NDVI, and RH using data collected at Earlsboro, OK (EF27) from 2003 to 2006.
Citation: Journal of Applied Meteorology and Climatology 48, 3; 10.1175/2008JAMC1959.1
An example of the ratio Rn/Sn as a function of DTaR, (Ta min), NDVI, and RH using data collected at Earlsboro, OK (EF27) from 2003 to 2006.
Citation: Journal of Applied Meteorology and Climatology 48, 3; 10.1175/2008JAMC1959.1
Scatterplot of measured and predicted daytime Rn calculated with Eq. (6) and daily minimum land surface temperature (Ta min), daily land surface temperature range (DTR), RH, and NDVI using data collected at Pawhuska, OK (EF12) from 2002 to 2006.
Citation: Journal of Applied Meteorology and Climatology 48, 3; 10.1175/2008JAMC1959.1
Scatterplot of measured and predicted daytime Rn calculated with Eq. (6) and daily minimum land surface temperature (Ta min), daily land surface temperature range (DTR), RH, and NDVI using data collected at Pawhuska, OK (EF12) from 2002 to 2006.
Citation: Journal of Applied Meteorology and Climatology 48, 3; 10.1175/2008JAMC1959.1
Scatterplot of measured and predicted daytime Rn calculated with Eq. (6) and daily minimum land surface temperature (Ta min), daily land surface temperature range (DTR), RH, and NDVI using data collected at Pawhuska, OK (EF12) from 2002 to 2006.
Citation: Journal of Applied Meteorology and Climatology 48, 3; 10.1175/2008JAMC1959.1
A comparison of the measured and predicted annual abnormality in daytime Rn calculated with Eq. (6), the coefficients listed in Table 2, and data collected at the 24 sites described in Table 1.
Citation: Journal of Applied Meteorology and Climatology 48, 3; 10.1175/2008JAMC1959.1
A comparison of the measured and predicted annual abnormality in daytime Rn calculated with Eq. (6), the coefficients listed in Table 2, and data collected at the 24 sites described in Table 1.
Citation: Journal of Applied Meteorology and Climatology 48, 3; 10.1175/2008JAMC1959.1
A comparison of the measured and predicted annual abnormality in daytime Rn calculated with Eq. (6), the coefficients listed in Table 2, and data collected at the 24 sites described in Table 1.
Citation: Journal of Applied Meteorology and Climatology 48, 3; 10.1175/2008JAMC1959.1
A description of the data measurement sites used in this study. Ratio Rn/Sn and RH are 30-min-average values. MODIS NDVI is the 16-day averaged product. We calculated the multiyear average values from 30-min-average values to characterize the climatological behavior of the sites.
A summary of the correlation coefficient between the ratio Rn/Sn and daily average air temperature (Ta daily), daily maximum (or minimum) air temperature (Ta max or Ta min), DTaR, daily average land surface temperature (Ts daily), daily maximum (or minimum) land surface temperature (Ts max or Ts min), DTsR, Sn, RH, NDVI, and EVI.
The fitted parameters in Eq. (6) and the statistics for all 24 sites. Equation (6) is used to predict daytime Rn at all 24 sites using the parameters shown in the following columns. The correlation coefficients and RMSE between measured and predicted daytime Rn are given in the last two rows.
A summary of the fitted statistics from Eq. (6) using daily minimum land surface temperature (Ta min), DTsR, RH, and NDVI. The same coefficients are used for all 24 sites.
