Potencial de la inteligencia artificial para avanzar en el estudio de la desertificación

Laburpena

Desertification is a global problem affecting 1.5 billion people living in the poorest and most vulnerable parts of the world. In recent years, several studies have contributed to provide information to assess the problem. Some of them are based on analysing biophysical and socio-economic variables using artificial intelligence techniques. For example, artificial intelligence have been used to complete anomaly data for water storage estimation, accurate identification of land cover, estimation of global daily solar radiation and improved climate predictions, and others. Although their use is not yet widespread, the future in desertification studies looks promising. In this paper we review the potential of artificial intelligence techniques (machine learning and deep learning) in the study of desertification and its recent growth in recent years. During the period 2015-2020 the number of publications implementing deep learning increased by 63%, while for machine learning its growth was more modest at 3%. In particular, when we look for studies related to desertification, the growth figures are more striking: an average increase of 103% in studies with deep learning, and 43% in machine learning. However, more studies and efforts are needed to bring together all the disciplines involved in the study of desertification in order to obtain a global and transversal vision of this phenomenon and thus design effective actions to mitigate or anticipate its adverse effects.