Para estimar con precisión la temperatura de la superficie terrestre (TST) es necesario conocer con exactitud la emisividad de dicha superficie. Sobre zonas vegetadas, la estimación de la emisividad es más compleja debido a las reflexiones múltiples de la radiación (emitida y reflejada) entre los distintos elementos que forman la vegetación (suelo y elementos de la planta). En la literatura pueden encontrarse diversos métodos que modelan la emisividad de la vegetación. En este trabajo, se ha profundizado en el análisis y evaluación de los modelos de transferencia radiativa (MTR) para la obtención de la emisividad direccional de la vegetación. Los modelos utilizados para este estudio han sido: FR97 (François et al., 1997), Mod3 (François, 2002), Rmod3 (Shi, 2011), REN15 (Ren et al., 2015), CE-P (Cao et al., 2018) y 4SAIL (Verhoef et al., 2007). Estos modelos han sido validados con medidas in-situ, aplicados a datos de satélite y evaluados cuando las emisividades resultantes se usan como entrada de un algoritmo split-window (SW) para la obtención de la TST. Validación de los MTR Los MTR fueron evaluados realizando medidas in-situ de emisividad sobre una muestra de rosales y dos suelos con características distintas: un suelo orgánico con alta emisividad y un suelo inorgánico (arena) con baja emisividad. Para la realización de medidas in-situ se utilizaron dos radiómetros CE312 con cinco canales estrechos entre 8 y 13 µm, los cuales permitieron aplicar un método de separación temperatura-emisividad (TES). En un primer estudio, donde se utilizó un suelo orgánico, se realizaron 15 medidas de radiancia para 7 ángulos de observación y 6 valores de LAI en cada caso. En un segundo estudio, el suelo orgánico se cambió por arena y se tomaron 15 medidas de radiancia para 2 ángulos de observación distintos y 6 valores de LAI. Además, para la aplicación de los modelos se midieron las muestras de hoja de rosal y de los dos suelos. De estas medidas se obtuvo una emisividad constante en el caso de las hojas, cercana a 0,98. Para el suelo orgánico, la emisividad obtenida varió entre 0,949 y 0,967 según el canal espectral, mientras que para el suelo inorgánico la emisividad mostró tener un mayor contraste espectral, con una emisividad entre 0,732 y 0,962. La emisividad derivada de los MTR se comparó primero con la emisividad medida en observación a nadir. Para el suelo orgánico, los resultados mostraron que las medidas de emisividad TES no diferían con la variación del LAI. Sin embargo, sobre el suelo inorgánico se observó en la emisividad TES el incremento en emisividad predicho por los MTR, principalmente debido a la mayor diferencia entre la emisividad del suelo inorgánico y de las hojas. Esta variación de la emisividad con el LAI, observada con las medidas in-situ, fue ajustada mediante regresión, obteniendo coeficientes de correlación entre 0,986 y 0,999 según el canal espectral. El MTR Mod3 obtuvo valores más cercanos a las medidas TES que el resto de modelos, teniendo en cuenta el análisis estadístico tanto del suelo orgánico como inorgánico. Estos mejores resultados del MTR Mod3 destacaron especialmente en aquellos canales donde la diferencia entre la emisividad de suelo y hoja era mayor. La evaluación de la emisividad obtenida en función del ángulo de observación mostró poca variación, tanto en el caso de los MTR como en el de las medidas TES, sobre cualquiera de las muestras con los distintos suelos utilizados. Por lo tanto, para condiciones similares a las analizadas en este estudio no se espera ninguna variación de la emisividad de la vegetación con el ángulo. En cuanto a la comparación entre los MTR y el método TES, para el suelo orgánico, donde la diferencia en emisividad entre suelo y hoja es mínima, el modelo Mod3 en global obtuvo los valores más cercanos a la emisividad TES en términos de RMSE. Sin embargo, estos valores fueron muy próximos a los obtenidos por los modelos FR97, REN15, CE-P y 4SAIL, con diferencias inferiores a la incertidumbre de las medidas. Si analizamos los resultados en función del intervalo de variación del LAI, el MTR que mejor se ajustó a los valores TES obtenidos de las medidas realizadas cuando LAI > 1,5 m2/m2 fue el MTR Mod3, pero para LAI < 1,5 m2/m2 el modelo con resultados más próximos a la emisividad TES fue el FR97. Cuando la diferencia entre la emisividad de suelo y hoja aumentó, como fue el caso de las medidas en el intervalo espectral 8,0 – 9,5 µm sobre la muestra con suelo inorgánico, el MTR Mod3 obtuvo los mejores resultados comparando con las medidas de emisividad TES para todos los valores de LAI. En este caso, los modelos FR97, REN15, CE-P y 4SAIL sobreestimaron considerablemente las medidas TES y el MTR Rmod3 la subestimó, especialmente para valores de LAI < 2 m2/m2. Aplicación a satélite Los MTR se utilizaron para generar mapas de emisividad direccional de la vegetación para imágenes del sensor MODIS a bordo del satélite EOS – Aqua sobre la Península Ibérica. El MTR Rmod3 fue excluido de este estudio debido a las grandes diferencias observadas en las medidas in-situ y su mayor coste computacional en la generación de los mapas. Estos mapas fueron generados para una rejilla de coordenadas constante con proyección WGS84 a una resolución espacial de 500 m, como máxima resolución, y a una resolución de 1 km para la comparación directa con los productos de emisividad MYD11 y MYD21 del sensor MODIS. Para generar los mapas de emisividad direccional de la vegetación se utilizaron distintos productos MODIS: el producto de LAI MCD15A2H, el producto de tipo de cobertura de superficie MCD12Q1 y el producto de temperatura y emisividad MYD21A1 (obtenido con el uso de un método TES), el cual fue utilizado tanto para obtener el ángulo de observación de cada pixel, así como para una posterior comparación con los datos obtenidos. Además, para la asignación de la emisividad de suelo, se utilizó un mapa mundial de clasificación de taxonomía de suelo. Las emisividades seleccionadas tanto para cada suelo como tipo de hoja fueron obtenidas a partir de los distintos espectros de la librería espectral ECOSTRESS (Meerdink et al., 2019). La emisividad obtenida al aplicar los MTR con datos de satélite fue comparada con datos de referencia tomados en una zona experimental de validación en Cortes de Pallás. Esta zona es una planicie de ~200 km2 de matorrales. Al tratarse de una zona con poca variación de la cantidad de vegetación a lo largo del año, no se observó ninguna variación temporal. En cuanto a los MTR, estos mostraron una diferencia entre ellos inferior a 0,015, siendo la mayor diferencia la observada entre los modelos CE-P (emisividad ~0,985) y Mod3 (emisividad ~0,970). Esta emisividad fue comparada con la obtenida por los productos MODIS MYD11A1, MYD11B1 y MYD21A1 y el producto de emisividad IREMIS de la Universidad de Wisconsin. Los valores de emisividad extraídos de los productos MODIS eran concordantes con los valores obtenidos con los MTR sobre la zona de estudio. El producto MYD21A1 mostró una mayor dispersión de datos, especialmente con valores de emisividad inferiores a los esperados. Esta dispersión pudo ser debida a la mayor sensibilidad en la corrección atmosférica del método TES utilizado para su generación. Tal como ocurrió en las medidas experimentales, para el suelo orgánico y valores de LAI cercanos a 1 m2/m2, el MTR FR97 se ajustó mejor a los valores obtenidos con los productos MODIS. Por otra parte, el producto IREMIS mostró valores de emisividad poco realistas, con valores propios de suelo sin vegetación, los cuales no coinciden con la descripción de la zona de estudio. La emisividad direccional de los MTR fue comparada frente al producto MYD21A1 sobre las distintas clases de coberturas de superficie de la Península Ibérica. El Mod3 mostró, en términos globales, un resultado más cercano a la emisividad obtenida con el producto MYD21A1 en los canales centrados en 11 y 12 µm. En cambio, en el canal centrado en 8,55 µm se observaron resultados similares en todos los modelos. Todos ellos obtuvieron mejores resultados en los canales de 11 y 12 µm, donde las diferencias entre la emisividad de las componentes de la superficie son, generalmente, más pequeñas. En el canal de 8,55 µm tanto la emisividad de suelo como de las hojas cubren un mayor intervalo, esto provocó una mayor dispersión en los datos observados, así como un mayor error sistemático en las distintas clases de superficie. Evaluación del efecto de la emisividad de los MTR sobre la TST La validación de los productos de TST es una tarea necesaria para el control del funcionamiento de los sensores térmicos a bordo de satélites, así como la mejora de la precisión de los algoritmos utilizados para la obtención de la TST. Con el fin de evaluar el efecto de la emisividad estimada a partir de cada MTR sobre la TST, la emisividad de los canales de 11 y 12 µm se utilizó en el algoritmo SW usado como operativo para MODIS (Wan, 2014), con los coeficientes propuestos por Wang et al. (2019). La TST obtenida con la emisividad de cada modelo, así como con la emisividad de los distintos productos MODIS, fue comparada con datos tomados in-situ en la zona de validación de Cortes de Pallás, tanto con los datos obtenidos con un sistema de barrido angular como con los radiómetros instalados con ángulo de observación fijo. Los resultados obtenidos de la comparación con ambos sistemas fueron similares. En los dos casos, se observó un error sistemático entre +0,0 y +0,6 K y un error aleatorio cercano a ±1 K para todos los MTR y productos MODIS, excepto para el producto de emisividad IREMIS, que obtuvo un error sistemático superior a +1 K. Dentro de estos resultados, la TST obtenida con la emisividad de los MTR CE-P, REN15, 4SAIL y FR97 obtuvo el menor error sistemático (< 0,2 K). En cuanto a la TST obtenida con la emisividad de los productos de MODIS MYD11A1, MYD11B1 y MYD21A1, mostraron resultados muy próximos a los obtenidos con los MTR, ya que la emisividad de estos productos es similar a la obtenida por los MTR en esta zona. La validación de la TST realizada para estos mismos productos sobre la zona obtuvo buenos resultados en términos de exactitud, con un error sistemático de +0,3 K y +0,4 K para los productos MYD11A1 y MYD21A1, respectivamente. Finalmente, se realizó complementariamente una validación R-based sobre dos zonas en la Península Ibérica: una de olivos y una de viña. En esta validación se observó un error aleatorio inferior a ±1 K en ambas zonas para la temperatura obtenida con la emisividad proporcionada por los MTR. En cuanto al error sistemático obtenido por los MTR fue distinto para cada una de las zonas, siendo entre -0,6 y -0,1 K en la zona de olivos y entre -0,1 y +0,2 K en la viña. Para la temperatura determinada con la emisividad del producto MYD11A1 se obtuvo un error sistemático y aleatorio igual que los MTR 4SAIL y REN15 sobre la zona de olivos y un error sistemático ligeramente superior en la zona de viña. Por lo tanto, se observaron mejores resultados con el MTR Mod3 que con la emisividad de los otros MTR y de los productos MODIS sobre la zona de olivos. En cambio, estos resultados fueron similares tanto en la zona de validación de Cortes de Pallás, con datos de validación in-situ, como sobre la zona de viña, con datos de validación R-based. En este último caso, a pesar del valor reducido de fracción de cobertura vegetal (< 0,27) y de LAI (< 0,7 m2/m2), la emisividad obtenida por el producto MYD11A1 era muy elevada, propia de una superficie vegetada (> 0,984). Sin embargo, la emisividad obtenida por los MTR (entre 0,97 y 0,98) resultaba más realista, teniendo en cuenta el tipo de superficie de la zona de estudio. Con todo ello, los MTR demuestran proporcionar buenos resultados en la estimación de la emisividad requerida para la determinación de la TST desde satélites con sensores térmicos como el MODIS, ya que los resultados son similares o, en algunos casos, incluso mejores a los obtenidos mediante los procedimientos actualmente operativos. Estos resultados hacen interesante la extensión de los MTR a otros sensores satelitales, para los que permitiría obtener la emisividad de la superficie considerando las posibles variaciones en la cubierta vegetal en términos de LAI.Introduction and objectives Land surface temperature (LST) was recognized as an essential climate variable by the World Meteorological Organization, as it is directly related with the energy balance between the Earth surface and the atmosphere. Atmospheric absorption and surface emissivity corrections are the main factors that affect an accurate retrieval of LST for data acquired from satellite sensors in the thermal infrared (TIR) spectrum. Therefore, an accurate characterization of the surface emissivity on the TIR spectrum is required for an accurate retrieval of the LST. The surface emissivity was well-characterized in the last decades for homogeneous and flat surfaces, e.g. water or arid bare soil sites. However, for heterogeneous surfaces the emissivity modeling is a more challenging point because of its structural complexity. The difficulties on the estimation of the canopy emissivity are higher due to the multiple reflections of the radiance (emitted and reflected) among the canopy components (soil and plant elements). For that, it is necessary to consider the multiple reflections that take place inside the canopy when the emissivity is being modeled. Different canopy emissivity models are found in the literature. These models can be classified as: geometrical models, bidirectional reflectance distribution function (BRDF) models and radiative transfer models (RTMs). In this work, several RTMs to obtain the directional canopy emissivity were analyzed and evaluated. These RTMs were: FR97 (François et al., 1997), Mod3 (François, 2002), Rmod3 (Shi, 2011), REN15 (Ren et al., 2015), CE-P (Cao et al., 2018) and 4SAIL (Verhoef et al., 2007). These models have common input parameters, which are the soil and leaf emissivity, the observation zenith angles and the leaf area index (LAI). The RTMs were validated against in-situ measurements and applied to satellite data. Moreover, they were used to obtain the LST applying a split-window (SW) algorithm, and the retrieved LST was evaluated with in-situ data. The main objectives of this thesis are: - To evaluate the RTMs performance over canopy with in-situ measurements, - evaluate the angular variation of the canopy emissivity from the RTMs and in-situ measured data, - generate directional emissivity maps using the RTMs for moderate resolution data (e.g., 500 m and 1 km), - compare the directional emissivity maps with the MODIS MYD21 emissivity product, which is obtained with the TES method, - and evaluate the RTMs emissivity effect when applying them to retrieve the LST from satellite data. Methodology The FR97 and Mod3 models are based on the RTM proposed by Prevot (1985) which takes into account the soil and leaves contributions to the canopy emissivity. The main difference between these models lies in the fact that the Mod3 models does not take into account the multiple reflections that takes place among the vegetated components (i.e. the different leaves inside the canopy), considering just the interaction between the soil and the leaves. For that, the cavity effect coefficient is not used in the Mod3 model, while it is part of the leaves contribution in the FR97 model. The Rmod3 model was presented as a modified version of the Mod3 model for satellite mixed pixels. This model introduced the vegetation fraction (Pv) as an input parameter modifying the Mod3 model, and added an additional term relating the bare soil emissivity with the bare soil fraction (1 – Pv). The 4SAIL model is an extension to the TIR of the four components scattering by arbitrarily inclined leaves (SAIL) model. The model is expressed in four differential equations which describe the interaction among four fluxes (two direct and two diffuse fluxes). A free distributed program code to solve these equations analytically was used in this study. The REN15 model follows the theory of the FR97 model, but it uses the 4SAIL model to estimate the cavity effect coefficient instead of the given coefficients for the FR97 model. Due to this modification, REN15 obtains closer values to the 4SAIL model than the FR97. The CE-P model is based on the recollision probability parameter instead of the cavity effect coefficient to consider the multiple reflections inside the canopy. This parameter is defined as the probability of a photon to interact with a canopy component after an interaction with another component. The recollision parameter was originally used in the visible and near infrared spectral range, and it is extended in this model to the TIR spectrum. A sensibility analysis of the RTMs was done to estimate each model uncertainty and the contribution of each parameter to the uncertainty. For that, typical soil and leaves emissivities were used, i.e. 0.94 and 0.98, respectively. It was estimated for observation zenith angles ranging from 0o to 60o in steps of 10o, and for LAI values ranging from 0.5 to 3.0 m2/m2 in steps of 0.5 m2/m2. Uncertainties of ±0.01 were assigned to each input emissivity, ±0.5o to the input observation zenith angle and ±23 % to the LAI uncertainty. An emissivity uncertainty between ±0.003 and ±0.010 was observed for most models, depending on the LAI and observation zenith angle. The highest contribution to the models uncertainty was the input emissivities, showing the soil emissivity the highest contribution when LAI < 1 m2/m2 and the leaves emissivity when LAI > 1 m2/m2. The contribution of the LAI uncertainty to the model uncertainty was also relevant for LAIs < 2 m2/m2. The contribution of the observation zenith angle to the models uncertainties was negligible. The RTMs were validated with in-situ emissivity measurements over a set of rose plants and two soils with different features: an organic soil (OS) with high emissivity and an inorganic soil (IS, sand) with low emissivity. The rose plants were selected because they form a continuous canopy, which allowed to control the structure when cutting off the leaves, and the steam was strong but fine enough to have a negligible contribution on the canopy emissivity. Moreover, this continuous canopy is in agreement with the type of canopy for which the RTMs are defined. Two CIMEL Electronique CE312 radiometers were used to take in-situ radiance measurements for five narrow channels in the 8 – 14 µm spectral range. In-situ directional emissivity measurements were obtained using the Temperature and Emissivity Separation (TES) method. These radiometers were calibrated in 2016 with a reference blackbody source during an international experiment in the framework of the Fiducial Reference Measurements for validation of Surface Temperature from Satellites (FRM4STS) project. They are regularly calibrated in our laboratory with a temperature variable Landcal P80P blackbody source, which was also calibrated in the 2016 FRM4STS experiment showing a root mean square error (RMSE) of ±0.05 K. Both radiometers (CE1 and CE2) were calibrated following the guidelines proposed by the Joint Committee for Guides in Metrology (JCGM, 2008), taking into account the random and systematic uncertainties. It was obtained a total uncertainty of ±0.15 K and ±0.12 K for CE1 and CE2 radiometers, respectively. The experiment took place at the Physics Faculty of University of Valencia, Spain. All measurements were taken at nighttime on cloudless days, in order to avoid shadow effects and to reduce the atmospheric absorption contribution in the methodology applied for the in-situ emissivities retrieval. In a first study, where the organic soil was used, 15 radiance measurements were taken for 7 observation zenith angles (from 0o to 60o in steps of 10) and 6 different LAI values (ranging from 0 to 2.8 m2/m2). In a second study, the organic soil was replaced by sand, and 15 radiance measurements were taken for 2 observation zenith angles (0o and 55o) and 6 LAI values (ranging from 0 to 3.3 m2/m2). The two soil samples and rose leaves were also measured at nadir observation in order to apply the RTMs. For each sample measurement, a simultaneous radiance measurement was taken over a high reflectance gold panel in order to estimate the sky radiance. To obtain the in-situ emissivity, the TES method was applied with the measured radiances previously corrected from the atmospheric contribution by subtracting the measured sky radiance. An emissivity close to 0.98 was obtained for the leaves for all CE312 channel. For the OS, the measured emissivity was between 0.949 and 0.967 depending on the spectral channel, while for the IS a higher spectral contrast was observed, with emissivity values between 0.732 and 0.962. In addition, the angular emissivity variation between nadir and 60o was measured, obtaining an emissivity decrease of 0.01 for the 10 - 12 µm spectral range and of 0.02 for the 8.0 – 9.5 µm spectral range for the OS. However, for the IS, it was observed a decreasing of 0.03 for the 10 - 12 µm and of 0.06 for the 8.0 – 9.5 µm. The measured at-nadir emissivities were used as input for the RTMs. The LAI was measured with Pocket-LAI Android app after cutting out the quantity of leaves (to reduce the LAI value) and previously to each set of radiance measurements. For each LAI value, 36 Pocket-LAI measurements were taken, considering their average as the corresponding LAI value for using as input parameter. This application has an associated uncertainty for broadleaf samples of ±23 %. After the validation of the RTMs with in-situ data, the RTMs were used to generate directional emissivity maps over the Iberian Peninsula for scenes of the MODIS sensor onboard the EOS – Aqua satellite. The Rmod3 RTM was excluded of this study due to the large differences observed with the in-situ measurements and its higher computational cost in the generation of the maps. These maps were generated for a constant grid in WGS84 at a spatial resolution of 500 m, as maximum resolution, but also at spatial resolution of 1 km for comparison with MYD11 and MYD21 emissivity products of MODIS sensor. In order to generate the directional emissivity maps, different MODIS products were used: LAI product MCD15A2H, land cover type product MCD12Q1, and temperature and emissivity product MYD21A1 (obtained using the TES method), which was used to obtain the observation zenith angle at each pixel, but also for comparison with the emissivity data obtained from the RTMs. These products were downloaded from the AppEEARS web tool, which allows to download level 3 MODIS data for a polygon of coordinates chosen by the user. Moreover, this web tool allows to change the scene projection from the sinusoidal MODIS projection original of the MODIS level 3 data to the wanted WGS84 projection (among others). Additionally, a world soil taxonomy classification based on the USDA taxonomy was used to assign the soil emissivity. This soil classification map is produced with data from the World Soil Information Service (WoSIS) and distributed by the International Soil Reference and Information Center (ISRIC). The selected emissivities for each soil and leaves were obtained from different spectra of the ECOSTRESS spectral library. Then, these spectra were convolved with the response functions of the MODIS TIR channels centered at 8.55 µm (channel 29), 11.03 µm (channel 31) and 12.02 µm (channel 32). The canopy emissivity derived from the RTMs with satellite data was compared with reference data taken on an experimental site for LST validation at Cortes de Pallás (Valencia). This site is an ~200 km2 shrubland plain, with a constant Pv throughout the year close to 0.5. Different emissivity MODIS products (i.e., MYD11A1, MYD11B1 and MYD21A1 MODIS products) and the IREMIS product of the University of Wisconsin were also compared with the RTMs emissivity at this site for the period from March to September, 2015. The emissivity for the three MODIS products are generated by means of different methods: MYD11A1 product is based on tabulated emissivities for each land cover obtained from BRDF methods; MYD11B1 is based on the day/night algorithm, which makes use of the daily and nightly satellite overpass to break the LST and emissivity indeterminacy; and the MYD21A1 is based on the TES method applied with the MODIS TIR channels 29, 31 and 32. The IREMIS product is based on the version 4 of the MYD11B2 product, which is a monthly average of the MYD11B1 product. Additionally, a comparison with the emissivities of the MYD21A1 product was carried out over the Iberian Peninsula. This comparison was taken for the different vegetation land covers of the International Geosphere Biosphere Program (IGBP) land cover classification. All pixels of the year 2015 over the Iberian Peninsula which were identified as cloudless by the MODIS cloud mask product were used. The IGBP classes related with forest (IGBP classes from 1 to 5) were grouped into a unique ‘Forest’ class, as well as shrubland classes (IGBP classes 6 and 7) and savanna classes (IGBP classes 8 and 9) were grouped into ‘Shrubland’ and ‘Savanna’ classes, respectively. Also, bare soil class was added to this analysis, as it could contain vegetation in some periods throughout the year. The validation of the LST satellite products is a task required to control the performance of thermal sensors, as well as to improve the accuracy and precision of the algorithms applied for the LST retrieval. With the aim of evaluating the effect of the RTMs emissivity on the LST retrieval, the emissivity obtained for channels centered at 11 and 12 µm was used to apply the current MODIS operational SW algorithm (Wan, 2014) with the coefficients proposed by Wang et al. (2019). The brightness temperatures given by the MODIS MYD021KM product were used to apply the SW algorithm. The LSTs estimated with the emissivity of each RTM and the emissivity of the different MODIS emissivity products were compared with in-situ data at the Cortes de Pallás validation site for the period from March to September, 2015. Data acquired with Apogee-121 wide band (8 – 14 µm) radiometers installed on a system with sweeping measurements at defined angles were used as reference for the validation, but also measurements with radiometers installed at the same station with a fixed viewing angle. These radiometers were calibrated at our laboratory with a Landcal P80P blackbody source and with a reference blackbody source of the National Institute of Standards and Technology during an international campaign organized by the Committee on Earth Observation Satellites in 2009. In both calibrations, in the international campaign but also in our laboratory, these radiometers showed uncertainties equal or better than ±0.2 K, which were in accordance with the uncertainty given by the manufacturer. A total of 55 match-ups were available for the validation with the radiometer installed on the angular system and 107 match-ups were available for the fixed view radiometers. The LSTs retrieved using the RTMs emissivity and the emissivity MODIS products were also validated over an olive orchard and a vineyard. As any in-situ ground data were available at these sites, a radiance-based (R-based) validation was carried out. The R-based technique uses the temperature of the channel centered at 11 µm as reference (T11) corrected from the atmospheric absorption and emissivity. Furthermore, a procedure to detect the most precise atmospheric profiles (ΔT Test) was required for this method. In the ΔT Test, the temperatures T11 are assumed as valid reference data when the difference between the T11 and the corrected temperature of the channel centered at 12 µm is lower than the uncertainty obtained for these differences. After selecting the valid data, a total dataset of 130 match-ups were available at the olive orchard and 139 match-ups were available at the vineyard. Results The RTMs were firstly validated against in-situ TES emissivities taken from the experimental measurements over the rose plants with OS and IS at the background. Emissivities derived from the RTMs were first compared with the TES emissivities measured at nadir observations. TES emissivities ranged from 0.975 from 0.985 for the OS sample, depending on the LAI and CE312 radiometer spectral channel, while for the IS sample, the emissivities ranged from 0.887 to 0.984. For the IS sample, the emissivities of the 10 – 12 µm channels were similar to the OS sample emissivities, due to the sand emissivity on that spectral range is similar to that of the OS. However, higher differences were observed on the 8.0 – 9.5 µm spectral range. TES emissivities from the IS sample were adjusted by regression with LAI to a non-linear function, obtaining regression coefficients of 0.986 and 0.999 depending on the spectral range. Similar emissivity values were found for FR97, 4SAIL, REN15 and CE-P, especially with the OS sample. For lower LAI values, Mod3 model showed similar values to FR97 model, but they differ with LAI, yielding to differences up to 0.013 for both soils. In comparison with the TES emissivities at nadir observation, if the overall results are analyzed, the Mod3 model showed the best agreement, with a RMSE of ±0.004 (±0.005) over the OS (IS) sample. However, over the OS sample, similar results were obtained with the other models, with RMSE values ranging from ±0.007 to ±0.009. Furthermore, for LAIs ≤ 1 m2/m2 the FR97, 4SAIL, REN15 and CE-P models showed better results than the Mod3. These models disagree in higher manner with the TES emissivities over the IS sample, where the RMSE values range from ±0.010 to ±0.013. In this case, the Rmod3 model obtained an overall RMSE of ±0.036, being up to ±0.06 for LAI = 0.5 m2/m2. The angular in-situ measurements were used to evaluate the angular performance of the RTMs, but no variation was found on the TES emissivities over the OS sample for the observation zenith angles measured in this study. Over the IS sample, very little variation was observed for LAIs ≤ 1 m2/m2, but it was within the emissivity uncertainty. In the comparison of the RTMs values with the TES emissivities, over the OS sample, the FR97, 4SAIL, REN15 and CE-P differences increased with LAI, with RMSE values ranging from ±0.002 to ±0.012. However, for the Mod3 and Rmod3 models, these differences decreased with LAI. In this case, the RMSE for the Mod3 models ranged from ±0.002 to ±0.007. If these results are analyzed according to the LAI range, for LAI > 1.5 m2/m2 the Mod3 model showed a better agreement with the TES emissivity, but for LAI < 1.5 m2/m2, the FR97 model showed closer emissivities to the ones obtained with the in-situ measurements. Over the IS sample, all models, except Rmod3, overestimated the TES emissivities. In this case, the Mod3 model showed the best agreement with the TES emissivities for all LAIs, with RMSE values ranging from ±0.002 to ±0.014, where the higher RMSE were found for the lower LAIs. The other RTMs showed RMSE values from ±0.009 to ±0.0018. Therefore, the Mod3 model was observed to obtain overall better results from the comparison with the in-situ TES emissivities over the rose plants samples. The FR97, Mod3, 4SAIL, REN15 and CE-P RTMs were used to generate directional emissivity maps over the Iberian Peninsula. These RTMs were compared with the MODIS emissivity products MYD11A1, MYD11B1 and MYD21A1 and the IREMIS product of the University of Wisconsin over the Cortes de Pallás LST validation site. As there is little variation on the quantity of vegetation over this site throughout the year, no significant temporal variation was observed. The differences observed among the RTMs were lower than 0.015, with the maximum emissivity difference between the CE-P model (~0.985) and the Mod3 model (~0.970). Just a little effect with the observation zenith angle was showed by the FR97 and Mod3 model, which was a periodically variation related with the revisit period of the MODIS – Aqua sensor. This emissivities were compared with the emissivity obtained by the MYD11A1, MYD11B1 and MYD21A1 MODIS products and the IREMIS emissivity product of the University of Wisconsin. The values extracted from the MODIS products were in agreement with the RTMs emissivity values at the study site, with values close to 0.980 in most cases. It was observed a higher variability for the MYD21A1 data, especially for the spectral channel centered at 8.55 µm, underestimating the expected values in many cases. This variability could be caused by the higher sensibility on the atmospheric correction of the TES method when it is applied to generate the product. As happened with the in-situ measurements for the case of OS and LAI values close to 1 m2/m2, the FR97 RTM obtained emissivity values closer to the emissivity obtained from the MODIS products. Contrarily, the IREMIS product showed unrealistic values, with emissivity values typical for bare soils (from 0.94 to 0.97), which are not consistent with the description of the surface at the study site. The directional emissivity maps were compared with the MYD21A1 emissivity product over the different surface types found at the Iberian Peninsula. In this case, due to the huge quantity of available data (more than 104 pixels for each class), robust statistics were used, i.e. median (systematic uncertainty), robust standard deviation (RSD, random uncertainty) and robust root mean square error (R-RMSE). From this comparison, FR97, 4SAIL, REN15 and CE-P overestimated the MYD21A1 emissivity for most classes and the three MODIS TIR channels, except for the shrubland class for channel centered at 8.55 µm and, in the case of the FR97 and REN15 models, for the savanna class for the same channel. The Mod3 model overestimated the MYD21A1 emissivity for the classes forest, savanna, shrubland and vegetation mosaics. The differences observed on the 8.55 µm channel were higher than at 11 and 12 µm channels, mainly due to the higher variability of the surface emissivity at the 8.55 µm channel. For channel centered at 11 µm, the random uncertainty for all models ranged from ±0.004 and ±0.005 and for the 12 µm channel ranged from ±0.006 to 0.007, being similar for all models. However, the median of the Mod3 was lower than for the other models, with an overall median value of +0.001 at both channels. For the FR97, 4SAIL, REN15 and CE-P models, these values ranged from +0.004 to +0.009 at 11 and 12 µm channels. The bare soil class showed larger systematic and random uncertainties for the 8.55 µm channel, with systematic (random) uncertainty values of +0.03 (±0.03) for all models. These large differences could be attributed to an incorrect estimation of the LAI which causes a high emissivity estimated by the models, an overestimation of the soil emissivity used as input parameter or a decrease of the soil emissivity with the observation angle which is not taken into account by the RTMs, since they are defined for canopy. The effect on the LST when the RTMs emissivity are applied to its estimation was evaluated applying the MODIS operational SW algorithm and using in-situ data acquired at the Cortes de Pallás validation site. Similar results were obtained from the comparison with the data acquired with the radiometers installed on the angular observation system and the radiometers installed with fixed viewing. In both cases, it was observed a systematic uncertainty between +0.0 and +0.6 K and a random uncertainty close to ±1 K for all RTMs and MODIS emissivity products, except for the IREMIS emissivity product, which showed a systematic uncertainty larger than +1 K. The FR97, REN15, CE-P and 4SAIL RTMs obtained the lowest systematic uncertainty (≤ 0.2 K). The LSTs retrieved with the emissivity of the MYD11A1, MYD11B1 and MYD21A1 products showed values in agreement with those obtained using the RTMs emissivity. This occurred because the observed MODIS and RTMs emissivities were similar at the site. The validation results of the LST at the site given by the MYD11A1 and MYD21A1 products showed a good performance, with a systematic uncertainty of +0.3 K and +0.4 K for the MYD11A1 and MYD21A1 products, respectively. Finally, a complementary R-based validation was applied over two sites in the Iberian Peninsula: an olive orchard and a vineyard. The estimated ΔT threshold to choose the most accurate atmospheric profiles was of ±0.7 K at the olive orchard and of ±0.5 K at the vineyard and the uncertainty of the reference T11 was of ±1.2 K at the olive orchard and of ±0.7 K at the vineyard. The validation results showed a random uncertainty lower than ±1 K at both sites for the LSTs retrieved with the RTMs emissivities. However, the systematic uncertainty was different in each site: it was between -0.6 and -0.1 K at the olive orchard and between -0.1 and +0.2 K at the vineyard. The Mod3 model showed a R-RMSE of ±0.8 K, while the other RTMs obtained a R-RMSE of 1.1 K. At the vineyard, the FR97 model obtained a R-RMSE of ±0.8 K, and it was of ±0.9 K for the other RTMs. The LST obtained using the MYD11A1 emissivity showed the same systematic and random uncertainty than the 4SAIL and REN15 RTMs at the olive orchard site, but the systematic uncertainty at the vineyard was slightly higher. The LST obtained with the MYD11B1 emissivity obtained a slightly higher random uncertainty than the RTMs at the olive orchard site, but the same results than the FR97 at the vineyard. The LSTs estimated with MYD21A1 emissivity showed a slightly higher random uncertainty at both sites. And the IREMIS product obtained a higher overestimation of the reference data, with a large systematic uncertainty. The LST given by the MYD11A1 and MYD21A1 products were also validated, showing a systematic and random uncertainty lower than 1 K at both sites. Conclusions The RTMs were validated with TES in-situ emissivity measurements over a set of rose plants with OS at the background in a first period and with IS at the background in a second period. The in-situ TES emissivities over the rose plants with OS showed no significant variation with LAI. Nevertheless, an increase of the emissivity with LAI was observed over the sample with IS, as it was predicted by the RTMs. This increase was mainly due to the significant difference between the IS and leaves emissivities. The observed variation of the emissivity with LAI was adjusted by regression, obtaining high correlation coefficients. For the overall results, the Mod3 RTM showed the best agreement with the TES emissivities. These results were especially better in the cases with higher differences between soil and leaves emissivities. From the evaluation with the viewing angle over the rose plants samples, little variation was observed for the emissivity derived from the RTMs but also for the measured emissivity. Then, for similar conditions to the ones analyzed in this study, no significant variation of the canopy emissivity with the viewing angle is expected. The Mod3 RTM obtained a slightly better agreement with TES emissivities over the sample with OS. But similar results were obtained by the FR97, REN15, CE-P and 4SAIL models. However, the Mod3 RTM showed the best agreement with the measured TES emissivities for all LAI values over the IS, especially for channels in 8.0 – 9.5 µm spectral range. In this case, the FR97, REN15, CE-P and 4SAIL models overestimated and the Rmod3 model underestimated the TES emissivity values, especially for LAI < 2 m2/m2. The RTMs were used to generate directional canopy emissivity maps. They were compared with MODIS emissivity products, i.e. MYD11A1, MYD11B1, MYD21A1 and IREMIS, at the Cortes de Pallás validation site. No variation throughout the year was observed in the RTMs emissivity because of the constant quantity of vegetation at the site. Just FR97 and Mod3 models showed a little variation of the emissivity with the observation zenith angle. The emissivity of the MYD11A1, MYD11B1 and MYD21A1 products were in agreement with the RTMs canopy emissivity. But the IREMIS product showed emissivity values lower than the expected, underestimating the emissivities obtained with the MODIS products and the RTMs. The directional emissivity maps produced with the RTMs were compared with the MYD21A1 emissivity product over the different surface types found at the Iberian Peninsula. In global terms, the Mod3 obtained the best agreement with the MYD21A1 emissivity for channels centered at 11 and 12 µm. However, similar results were obtained with all the RTMs for the channel centered at 8.55 µm. All the models obtained lower differences with the MYD21 emissivity for channels centered at 11 and 12 µm, where the differences between the emissivity of the canopy components are, typically, lower. The LSTs retrieved applying the MODIS operational SW algorithm with the emissivities from the directional emissivity maps were evaluated with in-situ data at the Cortes de Pallás validation site and with R-based data at an olive orchard and a vineyard. The results given by the Mod3 RTM showed the best agreement with the reference data at the olive orchard site. However, similar results were observed for all RTMs and MODIS emissivity products at the Cortes de Pallás validation site, using in-situ data as reference, as well as at the vineyard, using R-based validation data. For the latter, despite the low fraction vegetation cover (< 0.27) and LAI (< 0.7 m2/m2), the MYD11A1 emissivity was higher than expected, with typical values of full canopy surface (> 0.984). However, the RTMs emissivity values (between 0.97 and 0.98) were found more realistic according to the surface cover. These results showed that the RTMs provided good results in the estimation of the emissivity required for the LST retrieval from satellite thermal sensors, e. g. MODIS, since the results were similar or even better to those given by the currently operational procedures. The extension of the RTMs to other satellite sensors are shown to be interesting, as it would allow to estimate the canopy emissivity considering possible variations in the canopy in terms of LAI.