Comparison of in-situ, aircraft, and satellite land surface temperature measurements over a NOAA Climate Reference Network site Praveena Krishnan a,b, ⁎, John Kochendorfer a , Edward J. Dumas a,b , Pierre C. Guillevic c , C. Bruce Baker a , Tilden P. Meyers a , Borja Martos d a Atmospheric Turbulence and Diffusion Division, NOAA/ARL, Oak Ridge, TN, USA b Oak Ridge Associated Universities, Oak Ridge, TN, USA c Jet Propulsion Laboratory California Institute of Technology, Pasadena, CA, USA d The University of Tennessee Space Institute, Tullahoma, TN, USA abstract article info Article history: Received 27 January 2015 Received in revised form 17 April 2015 Accepted 16 May 2015 Available online xxxx Keywords: Land surface temperature Aircraft Satellite MODIS Validation USCRN Land surface temperature (LST) is a key variable for studying the energy and water vapor exchange at the bio- sphere–atmosphere interface. In an effort to better quantify the spatial variability and overall representativeness of single-point LST measurements being recorded at NOAA's Climate Reference Network (CRN) sites and to im- prove the accuracy of satellite LST measurements, airborne flight campaigns were conducted over a CRN site in Crossville, Tennessee, USA during 2010 to 2011. Multiple measurements of LST were made using infrared temperature sensors at micrometeorological tower sites and onboard an instrumented Piper Navajo airborne re- search aircraft. In addition to this, coincident LST products from the moderate resolution imaging spectroradiometer (MODIS) instruments (Collection 5), onboard NASA Terra and Aqua Earth Observing System satellites were used. In this paper the comparison of LST measurements made from multiple platforms are pre- sented. Our study showed that the temporal and spatial variability of surface temperature as indicated by the standard deviation of the brightness temperature (T b ) during the flight periods were b 1.7 °C. Aircraft and tower-based T b during the flight periods agreed well with a root mean square error (RMSE) of b 1.3 °C. The daytime MODIS LST was lower than the tower and aircraft-based LST, but higher than the daytime near surface air temperature (T a ). MODIS LST showed a positive and lower bias with the nighttime tower-based LST, but with slightly higher RMSE than the daytime dataset. The MODIS LST showed better correlation with the tower-based LST than T a during clear sky conditions due to the complex relationship between air and surface temperature. Including both day and nighttime data the MODIS LST showed a bias of -0.56 °C, RMSE of 2.84 °C, and standard deviations of the difference of 2.79 °C when compared to the mean tower-based LST at the site. Tower-based T a explained 98% of the variance in LST during nighttime conditions with a bias of ~0.8 °C and RMSE of 0.86 °C. © 2015 Elsevier Inc. All rights reserved. 1. Introduction Land surface temperature (LST) is a key variable in the study of the exchange of energy and water between the land surface and the atmo- sphere, and it influences land surface physical processes at local to glob- al scales (Sellers, Hall, Asrar, Strebel, & Murphy, 1988; Wan, Wang, & Li, 2004a). LST is the skin temperature of the land surface determined by radiometric measurements, i.e., the aggregate surface viewed by the sensor to a depth of about 12 μm(Norman & Becker, 1995), which can include vegetation, soil, and other land surface components (Betts, Ball, Beljaars, Miller, & Viterbo, 1996; Hall, Huemmrich, Goetz, Sellers, & Nickeson, 1992; Pinheiro, Mahoney, Privette, & Tucker, 2006). It is an indicator of the surface energy balance through its controls of the up- ward terrestrial radiation and surface fluxes (Pinker, Sun, Hung, Li, & Basara, 2009) and is influenced by surface-biophysical properties, vege- tation density and soil moisture. LST is critical for accurately retrieving important climate variables such as air temperature and relative humid- ity (Yao, Li, Li, & Zhang, 2011), soil moisture (Aires, Prigent, Rossow, & Rothstein, 2001; Wan, Wang, & Li, 2004a), canopy evapotranspiration (Wang & Liang, 2009; Wang et al., 2007) and surface heat and water fluxes (Anderson et al., 1997). It is assimilated into land surface models for short and medium range forecasting to improve land–atmosphere exchange simulations (Qin, Liang, Liu, Zhang, & Hu, 2007; Reichle, Kumar, Mahanama, Koster, & Liu, 2010; Rodell et al., 2004). Multi-year observations of LST have been applied to research topics such as assessing climate trends, observing extreme climate events (Jin, 2004), detecting land cover changes (French et al., 2008; Luyssaert et al., 2014); drought monitoring (Anderson et al., 2011; Rhee, Im, & Remote Sensing of Environment xxx (2015) xxx–xxx ⁎ Corresponding author at: Atmospheric Turbulence and Diffusion Division, NOAA/ARL, 456 South Illinois Avenue, Oak Ridge, TN 37830, USA. E-mail address: praveena.krishnan@noaa.gov (P. Krishnan). RSE-09422; No of Pages 16 http://dx.doi.org/10.1016/j.rse.2015.05.011 0034-4257/© 2015 Elsevier Inc. All rights reserved. Contents lists available at ScienceDirect Remote Sensing of Environment journal homepage: www.elsevier.com/locate/rse Please cite this article as: Krishnan, P., et al., Comparison of in-situ, aircraft, and satellite land surface temperature measurements over a NOAA Climate Reference Network site, Remote Sensing of Environment (2015), http://dx.doi.org/10.1016/j.rse.2015.05.011