Exploiting multisource remote sensing data for phenological monitoring in european forests and biodiversity conservation in tropical ecosystems
Author(s)
Cotrina Sánchez, Dany Alexander
Date Issued
July 18, 2025
Type
Doctoral Thesis
Abstract
Forests, on a global scale, play a fundamental role in regulating ecological processes and protecting biodiversity, especially in the context of constant climate change. In this context, phenological dynamics, particularly evident in temperate forests, are sensitive indicators of these changes. In contrast, tropical forests, recognized for their rich biodiversity, play a crucial role in regulating global climate patterns. However, forest cover loss and ecosystem degradation accelerate biodiversity decline and may push these ecosystems toward critical thresholds of ecological instability. To address these ecological challenges, remote sensing technologies, through a multiscale and multitemporal approach to land mapping, offer advanced tools to improve vegetation and species monitoring in various ecosystems. This thesis integrates remote sensors from emerging technologies, including terrestrial systems such as the 'TreeTalker©' (TT+), satellite multispectral sensors, and NASA's Global Ecosystem Dynamics Investigation (GEDI) mission's spatial Light Detection and Ranging (LiDAR), to address fundamental questions in ecology and conservation. Specifically, Chapter 1 presents a general introduction to passive and active remote sensors in terrestrial, satellite, and aerial systems, as well as their applications in forest ecology and conservation. Emphasis is placed on phenological monitoring in European forests and conservation efforts in tropical forests, which together define the study scope of this thesis. Chapter 2 presents the TT+ device, an Internet of Things (IoT)-based system designed to collect physiological parameters of trees. Its components and the network that enables data acquisition are detailed. Additionally, its applicability in species-level phenological monitoring in Mediterranean forests of Italy is explored. In Chapter 3, the Normalized Difference Vegetation Index (NDVI) obtained through the TT+ spectral sensor, which collects data below the canopy, is compared with the NDVI derived from Sentinel-2 satellite data in Fagus sylvatica plots across Italy. The advantages of combining both spectral data sources to improve forest phenological monitoring at a local scale and in different vegetation strata are highlighted. In addition, this chapter proposes improvements for data collection with the TT+ are proposed, emphasizing how the utilization of both data sources optimizes forest phenological monitoring. Expanding the spatial scale of analysis, in Chapter 4, GEDI LiDAR data are incorporated to evaluate its capability as an active sensor in detecting seasonal variations in the vertical structure of European forest canopies on a large scale. Using the Plant Area Index (PAI) from
iii
GEDI's L2B product, phenological metrics of start and end of season are identified and compared with those obtained from the Leaf Area Index (LAI) of the MODIS passive sensor. The similarities and differences found in the start and end dates of seasonality between GEDI-PAI and MODIS-LAI metrics are discussed, proposing an integrated approach. Finally, it is analyzed whether climatic variations prior to phenological events, such as air temperature and accumulated precipitation, influence the advancement or delay of phenological phases according to both sensors. Extending the geographic scope of analysis, in Chapter 5, the use of GEDI LiDAR data is extended in tropical forests of South America at large scale, specifically within Peruvian ecoregions, with a focus on biodiversity and conservation. The usefulness of vegetation structure metrics derived from GEDI is evaluated in the performance of species distribution models (SDM), and to monitor the status of forests, especially with respect to degradation inside and outside protected areas. Additionally, based on the SDM and auxiliary variables, priority areas were mapped to identify protected areas as 'core areas' that were interconnected through main ecological corridors via geospatial analysis. Core areas where deforestation rates over the past two decades were finally analyzed. These results are discussed in terms of the potential for integrating remote sensing data into conservation planning in Peru, and in the light of the potential improvements of Peru conservation strategy and plans. Finally, Chapter 6 presents the general conclusions, integrating the previous findings and highlighting their contributions to ecology and conservation. It also outlines future perspectives, emphasizing the potential of integrating data from multiple sources and scales to address challenges in forest monitoring and biodiversity conservation.
Additional information
Dottorato di ricerca in Scienze, Tecnologie e Biotecnologie per la Sostenibilità