Developing of new methods to forest heritage characterization and management through remote sensing

  1. Alonso Martínez, Laura
Zuzendaria:
  1. Juan Picos Martín Zuzendaria
  2. Julia Armesto González Zuzendaria

Defentsa unibertsitatea: Universidade de Vigo

Fecha de defensa: 2023(e)ko iraila-(a)k 20

Epaimahaia:
  1. Luis Barbero González Presidentea
  2. Raquel Díaz Vázquez Idazkaria
  3. Fabian Ewald Fassnacht Kidea
Saila:
  1. Enxeñaría dos recursos naturais e medio ambiente

Mota: Tesia

Laburpena

Information about forest resources is essential for the sustainable management of forests. This information is needed by a wide range of stakeholders, meaning a great variety in data requirements as well. Obtaining such a variety of data is a demanding task since the variables that need to be assessed are both complex and detailed. The information specifications regarding spatial and temporal resolutions are also challenging, especially under changing conditions such as climate change. Remote sensing is fundamental to cover this increasing need of data. The research on the usage of these technologies is advancing every day although challenges remain. Although remote sensing has been used to monitor different relevant variables at large scale (e.g. global, continental), there is a need to find methodologies for improving the use of these technologies for small-scale forest systems or areas continuously subjected to disturbances (both human and climate-driven). Research on monitoring certain tree species of relevance for forest stakeholders in this type of systems, as Eucalyptus sp. or specific formations of chestnuts forests, is also sparse. Finally, due to the increase in climate-driven disturbances, there is a need to better monitor natural disturbances and their impacts on this type of systems. In this dissertation we used different combinations of remote sensing sensors to address these challenges. In particular: Sentinel-2 images, WorldView-3 images, Aerial Laser Scanning, UAV-LiDAR, Terrestrial Laser Scanning and Portable Laser Scanning. We developed a methodology to generate a forest resources map at a landscape scale suitable for small-scale forest system with specific legend disaggregation needs (i.e., the need for mapping Eucalyptus spp. distribution) and dominated by forest disturbances. Besides, we developed methodologies to incorporate to the previous developed maps information on the location of disturbances and the identification of their drivers (particularly harvesting and wildfires). Additionally, we developed methodologies to automatically identify specific tree species (Pinus sp., Eucalyptus sp., Castanea sativa) and characterize their stands, again in areas dominated by small-scale forest systems. We also developed tools for the assessment and prevention of wildfires in this type of systems. We studied the impact of different environmental parameters on the development of large wildfires and we identified relevant areas for implementing preventing measures. In addition, we also developed a fuel model map operationally implementable on Atlantic landscapes. The conjunction of the methods developed to obtain forest resources maps that incorporate information on forest disturbances will allow for the continuous monitoring of forest resources in small-scale forest systems dominated by disturbances. Advances on monitoring species stands will help to inform decisions regarding key species of this type of forest systems. The studies focused on wildfires monitoring revealed that special attention should be paid to re-burned areas when planning wildfire management. Furthermore, we observed that cultural heritage elements have the potential to behave as fire management strategic areas and aid to increase the resilience of the system to wildfires. The fuel models map developed on this thesis are a key tool to inform decisions regarding wildfire prevention actions in Atlantic landscape. One of the most remarkable findings of this thesis is the strong potential of open-access data sources (Sentinel-2 and ALS) to monitor forests, and particular tree species stands, even in areas with small-scale systems and areas continuously subjected to disturbances. From an operational standpoint, this use of open-access data sources will mean that the products developed can be more easily implemented. All the methodologies developed in this thesis were published in international journals (JCR) or presented at international or national conferences. Some outcomes of this thesis have also helped to fulfil some national and international project needs. Additionally, a substantial effort has been made to translate the findings from these studies into practical forestry applications. A successful example of this effort is that some of the products developed are available to the public on the web page of the Galician regional administration of rural areas.