Browsing by Subject "Random Forest"

Now showing 1 - 4 of 4
  • Thumbnail Image
    Publication
    Open Access
    Distribution and behaviour of striped dolphins in the southwestern Mediterranean Sea based on whale-watching data
    (Pergamon-Elsevier Science Ltd, 2023-10-31) Canales Cáceres, Rosa; Gomariz Castillo, Francisco; Alonso Sarria, Francisco; Abel, Isabel; Gimenez Casalduero, Francisca; Geografía
    The striped dolphin (Stenella coeruleoalba) is a cosmopolitan cetacean and the most commonly sighted dolphin in the Mediterranean Sea. It usually appears in groups of very different sizes, ranging from less than ten to more than 500 individuals, although it is usually found in groups of between 21 and 50 individuals. In the western Mediterranean, and more specifically in the Gulf of Mazarrón, S. coeruleoalba was the most frequently sighted cetacean during the 1042 whale-watching trips. The goal of this study was to establish the spatial and temporal distribution of striped dolphin sightings along the Gulf of Mazarrón between 2004 and 2014. Spatial patterns were analysed using a Random Forest based Species Distribution Model to estimate the presence of the species. Twentythree variables (three geographic, one temporal, eight geomorphometric and twelve oceanographic) were used as predictors. Out of the 1042 cruises, 872 records of striped dolphins were obtained. Some variations in the grouping patterns of these mammals were observed during the years 2006–2007, with an average shift in the size of the groups to fewer individuals (3−10). This variation is probably related to an epizootic event of morbillivirus occurring during those years, which was responsible for an abnormal rate of strandings of striped dolphins and long-finned pilot whales (Globicephala melas). The Random Forest model allowed to select 6 predictors related to morphometry and sea currents, suggesting the importance of specific habitat in offshore areas between 1000 and 3000 m depth in the continental slope
  • Thumbnail Image
    Publication
    Open Access
    Estimation of Actual Evapotranspiration using TDTM model and MODIS derived variables
    (Taylor & Francis LTD, 2022-01-10) Ruiz-Álvarez, Marcos; Gomariz-Castillo, Francisco; Alonso-Sarria, Francisco; López Ballesteros, Ana; Geografía
    Abstract: The objective of this paper is to contribute to improve ETa estimation in semiarid environments by proposing two variations to the TDMT model. These variations are based on the use of MODIS products from TERRA or AQUA satellites and on the use of NDVI instead of EVI, to estimate the fraction of vegetation cover. The proposed changes were validated with the original methodology for the 2012-2014 period with data obtained from two flux towers (ES-LJu and ES-Amo). The best results were obtained when using the alternative methodology (RMSE of 0.64-0.67 mm in ES-Lju and of 0.97-1.02 mm in ES-Amo). Additionally, a correction of the temporal systematic errors of the model based on Random Forest is proposed. With this correction, RMSE of 0.31-0.35 mm in ES-Lju and 0.30-0.34 mm in ES-Amo were obtained. The spatial distribution obtained with this corrected model is the most consistent with the characteristics of the study area
  • Thumbnail Image
    Publication
    Open Access
    Modification of the random forest algorithm to avoid statistical dependence problems when classifying remote sensing imagery
    (Pergamon-Elsevier Science Ltd, 2017) Cánovas García, Fulgencio; Alonso Sarria, Francisco; Gomariz Castillo, Francisco; Oñate Valdivieso, Fernando; Geografía
    Random forest is a classification technique widely used in remote sensing. One of its advantages is that it produces an estimation of classification accuracy based on the so called out-of-bag cross-validation method. It is usually assumed that such estimation is not biased and may be used instead of validation based on an external data-set or a cross-validation external to the algorithm. In this paper we show that this is not necessarily the case when classifying remote sensing imagery using training areas with several pixels or objects. According to our results, out-of-bag cross-validation clearly overestimates accuracy, both overall and per class. The reason is that, in a training patch, pixels or objects are not independent (from a statistical point of view) of each other; however, they are split by bootstrapping into in-bag and out-of-bag as if they were really independent. We believe that putting whole patch, rather than pixels/objects, in one or the other set would produce a less biased out-of-bag cross-validation. To deal with the problem, we propose a modification of the random forest algorithm to split training patches instead of the pixels (or objects) that compose them. This modified algorithm does not overestimate accuracy and has no lower predictive capability than the original. When its results are validated with an external data-set, the accuracy is not different from that obtained with the original algorithm. We analysed three remote sensing images with different classification approaches (pixel and object based); in the three cases reported, the modification we propose produces a less biased accuracy estimation.
  • Thumbnail Image
    Publication
    Open Access
    Obtención de cartografías de usos y coberturas del suelo de la Demarcación Hidrográfica del Segura para el periodo 1986-2019, empleando teledetección y clasificación digital de imágenes.
    (Universidad de Murcia, Servicio de Publicaciones, 2021) Rodríguez Valero, María Isabel; Alonso Sarriá, Francisco
    Los cambios en los usos y coberturas del suelo conllevan consecuencias ambientales de diverso tipo. La clasificación digital de imágenes obtenidas mediante teledetección es una potente herramienta para evaluar el grado de transformación ambiental debido a factores antropogénicos. El objetivo de este trabajo es desarrollar un esquema de trabajo basado en técnicas de teledetección y clasificación digital de imágenes para obtener cartografías de usos y coberturas del suelo en la Demarcación Hidrográfica del Segura, para el periodo comprendido entre 1986 y 2019. Para ello se ha utilizado el algoritmo de clasificación supervisada Random Forest y como variables predictoras las bandas espectrales de imágenes de Landsat 5, 7 y 8, y cuatro índices normalizados derivados de estas. Aunque la exactitud global conseguida indica que existe un margen de mejora en el ajuste del modelo, las clasificaciones obtenidas se consideran fiables. Respecto a los usos y coberturas del suelo obtenidos tras el proceso de clasificación, se observa una disminución del uso forestal y un aumento de los usos agrícolas, la superficie cubierta por matorral, los suelos desnudos y los invernaderos.