Validation of remotely sensed surface temperature over an oak woodland landscape — The problem of viewing and illumination geometries
Introduction
Land Surface Temperature (LST) is an important climatological variable (Sellers, Hall, Asrar, Strebel, & Murphy, 1992) as well as a diagnostic parameter of land surface conditions. It plays an important role in the surface energy balance, and as such it has long been used to infer surface heat fluxes (Caparrini et al., 2004, Mannstein, 1987), soil moisture (Carlson, 1986, Nemani et al., 1993), evapotranspiration (Kustas & Norman, 1996) and vegetation properties (Lambin & Ehrlich, 1997), including vegetation hydric stress (Jackson, Idso, Reginato, & Pinter, 1981).
Remote sensing constitutes the most effective method to observe LST over large areas and on a regular basis. Satellite LST products generally rely on measurements within the atmospheric window in the thermal infrared (e.g., Li et al., 2013). As such, remote sensing retrievals of LST correspond to the directional radiometric temperature of the surface within the field of view of the sensor (e.g., Norman & Becker, 1995). The validation of LST retrievals is however not trivial, given its high variability in space and time, along with the anisotropic effects. Validation exercises are commonly performed through comparisons of LST against ground-based measurements or through a radiance-based method (e.g., Wan & Li, 2008). The latter involves using radiative transfer calculations to reconstruct top-of-atmosphere observations from the LST retrievals and assuming surface emissivity and atmospheric profiles are known. The former is usually performed over homogeneous areas such as lakes, deserts and dense or very homogeneous vegetation covers, where station measurements are representative of pixel scale values (Göttsche et al., 2013, Wan et al., 2002, Wan et al., 2004). For heterogeneous surfaces, however, validation can be much more complex as an effective upscaling of the ground measurements is needed (Guillevic et al., 2012).
The comparison of LST estimations obtained from sensors on-board different platforms provides useful insight on product consistency (e.g., Jiménez et al., 2012, Trigo, Monteiro, Olesen and Kabsch, 2008). There are, however, many possible sources of LST differences, and it is difficult to ascertain the actual accuracy of each retrieval. Discrepancies between LST products may be associated to differences (i) in the top-of-atmosphere measurements (sensor calibration, spatial resolutions), (ii) in the algorithm and auxiliary data used for atmospheric and surface emissivity correction, (iii) in cloud mask, and (iv) in angular anisotropy (e.g., Barroso et al., 2005, Pinheiro et al., 2006, Rasmussen et al., 2010). Furthermore, remotely sensed LST is a directional variable, unless some sort of compositing of observations from different viewing angles is performed. As such, hypothetical LST retrievals obtained for the same scene, using the same sensor, but at different viewing angles would likely produce different temperature values, depending on factors like surface type, soil characteristics and slope orientation relative to sun. Although surface structure exerts an important role on the temperature, due in particular to shadowing effects that result in a dependence of LST on the zenith and azimuth view angles, these effects are often disregarded. In validation exercises involving comparisons of LST estimations with in situ observations, or inter-comparisons of LST products, the viewing and illumination geometries should be taken into account.
The effects of viewing and illumination geometries are usually considered by means of geometrical–optical models that have been developed mainly to describe forests and other discontinuous canopies. They operate by assuming that the canopy may be described by an array of geometrical objects arranged in space according to some statistical distribution. The interception and reflection of radiation are computed analytically from geometrical considerations. For these models, the overall radiance at any angle is calculated as a weighted average of the radiances from each component (usually, sunlit and shaded background and sunlit and shaded canopy).
This study presents a geometrical model that allows estimating the projected areas of the different components using parallel-ray geometry to describe the illumination of a three-dimensional vegetation element and the shadow it casts. The proposed model not only allows the correction of LST differences between sensors associated with their viewing geometries, but it is also an effective means for the validation of satellite-derived LST with ground-based measurements.
This type of geometric-optical model has been used by several authors to solve radiative transfer problems associated with surface heterogeneities related to vegetation (Franklin and Strahler, 1988, Lagouarde et al., 1995, Li and Strahler, 1986, Li and Strahler, 1992, Ni et al., 1999, Strahler and Jupp, 1990), as well as in studies of surface temperature anisotropy (Minnis and Khaiyer, 2000, Pinheiro et al., 2006, Rasmussen et al., 2010, Guillevic et al., 2013). Instead of relying on a rigid analytical approach, the procedure developed here has the advantage of using a simple computational method to calculate the geometrical projections, while making very few a priori assumptions. The method consists of projecting a three-dimensional vegetation object onto a fine grid, which allows the use of any vegetation shape and size or the combination of different shapes and sizes.
The model is applied to in situ measurements of brightness temperature gathered at Évora validation site to obtain the ground temperature corresponding to any observation and illumination angles. The site is located in a region dominated by sparse canopies of evergreen oak trees (Southern Portugal; Kabsch, Olesen, & Prata, 2008). The resulting temperature is compared against LST data as obtained from the Spinning Enhanced Visible and Infrared Imager (SEVIRI) onboard Meteosat Second Generation (MSG) satellites (Trigo et al., 2011) and from the MODerate resolution Imaging Spectroradiometer (MODIS) onboard AQUA and TERRA (Salomonson, Barnes, & Masuoka, 2006). Finally, the geometric model is used together with in situ measurements to estimate and remove the LST differences between MSG and MODIS associated with the different viewing geometries.
Section snippets
Data and methods
This study concerns the validation of satellite LST products with in situ measurements collected at Évora validation site in Southern Portugal. The period under analysis spans from October 2011 to September 2012, although the data are limited to clear sky observations. All comparisons with ground data from Évora are for the LST estimations for the satellite pixel nearest to the station.
Model sensitivity to input parameters
The geometric model was used to perform a sensitivity study of in situ LST to the parameters listed in Table 1. Results show that the impact is highest for daytime observations during summer months (June–September). An increase/decrease of 5% in PTC would lead to cooling/warming of daytime LST of up to 1 °C between June and August and up to 0.5 °C in April–May. The impact is very small for the remaining months or at night-time. The variability in the canopy size has significantly lower impact
Satellite inter-comparison
The developed model may be also used to compare LST retrievals from different satellites. Here we compare MSG LST against MODSW (i.e., MOD11A1 and MYD11A1) LST retrievals overlapping in time over Évora. The dependence of the daytime LST differences between MSG and MODSW on MODIS viewing geometry has already been analyzed by Trigo, Monteiro, Olesen, and Kabsch (2008) and Guillevic et al. (2013). In order to further understand this effect, Fig. 8 displays the discrepancies between the two
Concluding remarks
This paper analyses the problem of comparing satellite estimations of Land Surface Temperature with in situ measurements taken at regions with sparse canopies, where we often have strong temperature differences between sunlit background, shaded background and tree crowns. For that purpose, we developed a procedure that allows estimating the impact of viewing and illumination geometries on LST observations from space, while at the same time keeping assumptions at a minimum. The methodology is
Acknowledgments
This work was carried out within the context of the LSA SAF project (http://landsaf.ipma.pt) funded by EUMETSAT (LSA SAF: CDOP-2 proposal).
References (49)
- et al.
Constraints on transpiration from an ever green oak tree in southern Portugal
Agricultural and Forest Meteorology
(2004) - et al.
Land Surface Temperature validation using NOAA's surface climate observation networks — Scaling methodology for the Visible Infrared Imager Radiometer Suite (VIIRS)
Remote Sensing of Environment
(2012) - et al.
An experimental study of angular effects on surface temperature for various plant canopies and bare soils
Agricultural and Forest Meteorology
(1995) - et al.
Land-cover changes in sub-Saharan Africa (1982–1991): Application of a change index based on remotely sensed surface temperature and vegetation indices at a continental scale
Remote Sensing of Environment
(1997) - et al.
Satellite-derived land surface temperature: Current status and perspectives
Remote Sensing of Environment
(2013) - et al.
Terminology in thermal infrared remote sensing of natural surfaces
Agricultural and Forest Meteorology
(1995) - et al.
Modeling bidirectional reflectance of forests and woodlands using Boolean models and geometric optics
Remote Sensing of Environment
(1990) New refinements and validation of the MODIS Land-Surface Temperature/Emissivity products
Remote Sensing of Environment
(2008)- et al.
Validation of the land-surface temperature products retrieved from Terra Moderate Resolution Imaging Spectroradiometer data
Remote Sensing of Environment
(2002) - et al.
Intercalibration of NOOA and Meteosat window channel brightness temperatures
International Journal of Remote Sensing
(2005)