Study of growth-environment relationships and optimisation of management including climatic uncertainty of radiata pine stands in Galicia
| Autor: | González Rodríguez, Miguel Ángel |
|---|---|
| Přispěvatelé: | Diéguez Aranda, Ulises, Crecente Campo, Felipe, Universidade de Santiago de Compostela. Escola de Doutoramento Internacional (EDIUS), Universidade de Santiago de Compostela. Programa de Doutoramento en Enxeñaría para o Desenvolvemento Rural e Civil |
| Rok vydání: | 2021 |
| Předmět: |
Investigación::31 Ciencias agrarias::3106 Ciencia forestal::310608 Silvicultura [Materias]
Investigación::53 Ciencias económicas::5312 Economía sectorial::531201 Agricultura silvicultura pesca [Materias] stand growth modelling stand-level man-agement climate change machine learning radiata pine site index Investigación::12 Matemáticas::1209 Estadística::120914 Técnicas de predicción estadística [Materias] risk modelling |
| Zdroj: | Minerva. Repositorio Institucional de la Universidad de Santiago de Compostela instname |
| Popis: | Climate change is intended to impact forest dynamics significantly inthe following decades. To proactively adapt forest management to these expected alterations, new methodologies for handling the uncertain-ties regarding forest growth under varying environmental conditions become necessary. The purpose of this thesis was to forecast the impact of climate change on radiata pine plantations in the northwest of Spain in terms of productivity, profitability, and silvicultural treatments. In Study I, several statistical techniques were used for predicting thesite index (SI) of radiata pine stands using environmental predictors extracted from available raster maps. A non-linear technique, Multivariate Adaptive Regression Splines (MARS), was suggested as the best modelling alternative, explaining up to 52% of the SI variability. In Study II, the Support Vector Regression technique was used to predict SI and delimit the validity area of predictions based on the radial basis kernel. The resulting model had high predictive performance, provided robust predictions under varied climatic conditions, and included a relatively small number of predictors. Moreover, the model was able to identify areas where climatic conditions were very different from the observed and consequently regularised predictions for those areas. In Study III, silviculture under climate change was optimised for maximising the soil expectation value of a set of radiata pine plantations. The future forest productivity projections, produced by the model developed in Study II, forecasted an overall reduction in SI under climate change, mainly driven by increased temperatures and continentality. Consequently, the economic simulations forecasted a drop in profitability under climate change that was more intense for more pessimistic scenarios (RCP 6.0). However, the climatic projections were very varied over the set of used climate models, which led to a great dispersion in productivity and profitability predictions. From the perspective of silviculture, the most notable forecasted variation is the expected increase in optimum rotation lengths. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|