Modelamiento correlativo de la distribución potencial de ecosistemas de la provincia de Loja periodo 1970-2080: potenciales cambios bajo un contexto de cambio climático

Climate change is an undeniable phenomenon, it has been identified as the main threat to the integrity and functionality of terrestrial ecosystems in the last century. In the present investigation, the terrestrial ecosystems of the province of Loja were modeled for the year 2080, seeking to determine their current and future potential distribution under a climate scenario of average carbon dioxide emissions IPCC - AR5 - RCP 6.0. The models were implemented in R using the Biomod2 library that generates consensus models from 10 state-of-the-art modeling techniques. Biomod2 consists of 3 important stages such as: calibration, evaluation and modeling/projection and requires records of presence and absence of ecosystems, as well as a set of 19 climatic and 4 topographic variables. In order to eliminate multicollinearity and overfitting of the models, two statistical methods were implemented: Pearson correlation and Variance Inflation Factor. To evaluate the performance of the models, two metrics based on machine learning were used, such as ROC and TSS, a threshold of 0.8 is considered as a metric to evaluate the models, the models that exceed this threshold are considered for the final assemblies and the badly evaluated will be discarded. The results indicate that the temperature would suffer an increase for all ecosystems, on average 3.2 °C; while precipitation would increase by an average of 110.5 mm, but not for all ecosystems, some ecosystems of altitudinal floors such as Páramo and Bosque Montano Alto would decrease. Climate variability possibly generates new climatic configurations and the ecosystems in response tend to redistribute, either gaining, losing or staying in their current distribution range. Keywords: Climate change, Modeling, Biomod2, Ecosystems, Loja.

• 第一著者: Poma Sarango, Diego Fernando (author)
• フォーマット: bachelorThesis
• 言語: spa
• 出版事項: 2023
• 主題: CAMBIO CLIMATICO; ECOSISTEMA; MODELAMIENTO
• オンライン･アクセス: https://dspace.unl.edu.ec/jspui/handle/123456789/27833