Titolo della tesi: Climate change and co-occurring disturbance effects on the Mediterranean forest ecosystems undergone to several management and socio-economic conditions.
The following PhD project is part of the NewLife4Drylands project (Remote Sensing-Oriented Nature-Based Solutions towards a New Life for Drylands) (Grant Agreement No. LIFE20 PRE/IT/000007). The main objective of the LIFE project is to monitor the application, scalability, and replicability of Nature-Based Solutions (NBS) for restoring degraded and arid areas using remote sensing data and modelling.
Within the scope of the LIFE project, this PhD research aimed to propose new tools and different techniques to assess the effects of climate change and aridity on Mediterranean ecosystems, particularly concerning natural area management.
Since the first assessment of climate change conducted by the United Nations and published by the Intergovernmental Panel on Climate Change (IPCC), hundreds of reports and scientific articles have described the past and potential impacts of climate change on species and their habitats. These studies have highlighted the complexity of predicting such changes and their effects on ecosystems. However, it is well-known that the availability of high-frequency data necessary for these studies poses a limitation in understanding the effects of climate changes, especially in highly heterogeneous areas. The Mediterranean Basin region falls into this category, and its characteristics make it particularly sensitive to the effects of climate change and human activities.
Based on these premises, this PhD work is structured around three main steps: optimizing the use of satellite data in the Mediterranean Basin area for studying historical series; implementing statistical analyses to define the aridity trends along the longitudinal transect of the Mediterranean studying climatic and vegetation trends in three study areas: Palo Laziale, Nestos, and Alta Murgia; making site-specific future predictions of climatic variables and defining changes in aridity trends and seasonality.
To optimize the use of satellite imagery for studying historical data in the Mediterranean Basin area, this study aimed to evaluate differences among various sensors and propose harmonization methods. This involved the integration of Landsat-7, Landsat-8, and Sentinel-2 surface reflectance products using Ordinary Least Squares (OLS) and Reduced Major Axis (RMA) regression models (Chapter 1). Analysing six years of spectral images and 1.500 random and evenly distributed samples in the Mediterranean Basin area, significant differences between sensor pairs were identified. Regression models (OLS and RMA) were introduced to harmonize these disparities, suggesting a mixed-band adjustment approach based on integrated coefficients from both models. Applying these coefficients significantly enhanced sensor harmonization, also confirmed by the Normalized Difference Vegetation Index (NDVI) calculations. This study provides a harmonization procedure, enabling comparable surface reflectance data for the Mediterranean Basin area. By using a combination of OLS and RMA regression models, the study underscores the importance of site-specific transformation coefficients, enhancing satellite data integration accuracy.
The obtained harmonization coefficients allowed the calculation of time series of remote sensing indices (NDVI, EVI) employed for an integrated analysis with the climatic variables (Chapter 2). An in-depth exploration was conducted to assess the impact of increasing aridity on vegetation in Palo Laziale, Nestos, and Alta Murgia. Utilizing advanced techniques such as Bayesian modeling (BEAST) and Cross-Wavelet Coherence (WCA), the study delved into the complex interplay between aridity and vegetation dynamics.
The findings revealed a concerning trend: over the last decade, these Mediterranean regions experienced a significant rise in aridity levels, impacting sites differently. Specifically, aridity intensified notably in Palo Laziale during winter (January-March) and autumn (October-December), whereas Nestos and Alta Murgia faced pronounced arid periods in the summer months of July, August, and September. This escalating aridity was quantified by the Standardized Precipitation Evapotranspiration Index (SPEI), calculated using both potential evapotranspiration (ET0) for SPEIpot and actual evapotranspiration (ETa) for SPEIact. Comparing the two indices highlighted differences in their impact on vegetation classes. The study emphasized the critical role of Shrubland vegetation in adapting to arid conditions, displaying remarkable resilience, particularly concerning the spread of rapidly expanding alien shrub species.
Through Bayesian Estimator of Abrupt Change, Seasonality, and Trend (BEAST) analysis and Cross-Wavelet Coherence Analysis (WCA), the intricate relationship between aridity and vegetation health in Mediterranean regions was explored. The integrated use of these analyses identified change points in vegetation and aridity indices, investigating the correlation between the two variables. WCA characterized this correlation in an innovative way, demonstrating its effectiveness in defining the periodicity and temporal lag characterizing aridity's influence on vegetation health. These findings, revealing complex relationships and temporal gradations, offer valuable insights into the impact of aridity on Mediterranean forest ecosystems.
The results highlighted various aspects of the studied sites, recognizing a significant seasonality in the influence of atmospheric water availability on vegetation during spring/summer quarters for Palo Laziale and Alta Murgia sites. Additionally, the study identified the years and conditions of aridity characterizing the Nestos site. Crucially, this study developers the integration of BEAST and WCA, demonstrating their potential in capturing the real dynamics of examined time series. While BEAST usage is increasing, the use of WCA, especially in the Mediterranean context, remains limited.
The outcomes of this study underscore the urgency of addressing escalating aridity in the Mediterranean region, emphasizing the need for site-specific analyses. The advancement of aridity in the last decade raises concerns about the future trends of climatic variables. Considering that local climatic variables exhibit specific patterns, localized prediction scenarios are crucial for quantifying the advancement of climate change and aridity in the most specific manner possible. The implementation of SARIMA models provided insights into the future trends of aridity (Chapter 3).
In the period 2023-2032, the Palo Laziale area will experience a minimum temperature increase of 0.75°C, aligning with IPCC predictions, while maximum temperatures will slightly rise. Precipitation will drastically decrease, particularly in the second and third quarters, causing a 29% increase in extremely arid months compared to 2000-2022. Aridity will extend to the third quarter from 2028 to 2032. Similar trends are observed in the Alta Murgia area. Despite no increase in aridity for the Nestos site, springtime humid quarters will significantly decrease. Thus, the importance of considering transitions from humid to dry conditions is emphasized for assessing the health of Mediterranean forests. The integration results emphasize advancing aridity in three representative study sites along a longitudinal transect of the Mediterranean, underscoring the necessity for studies in this direction.