Here mostly for historical reasons, I am now working on scientific research in my spare time. I still consider myself a professional (academic) scientist, although I might have transitioned into being a (skilled) dilettante.
Developing and applying species distribution models, with emphasis on the integration between mechanistic and correlative models;
Catastrophes such as storms, flood, droughts, diseases or invasion by a new competitor or predator occur widely, and they are usually driven by complex dynamics occurring at different spatial and temporal scales. Although these large unexpected changes may happen infrequently, they tend to dictate the long-term dynamical behavior of real-world natural systems, by either killing organisms, reshaping the environment, or creating new ecosystem dynamics.
Several investigations have explored the consequences of episodes of massive mortality on population dynamics and population persistence across several taxa. However, the vast majority of studies to date explored neither the adaptive mechanisms helping population recovery after collapses nor the role of catastrophes in shaping the life histories of the affected organisms and species. Given the predicted intensification of weather extremes and their increasing temporal autocorrelation with global climate change, the complex dynamics associated with adaptation and responses to catastrophic events call for long-term, global scientific investigations.
Marble trout, which I have used as a model system in this line of research, live in harsh habitats that are frequently affected by flash floods and debris‑flow causing massive fish mortalities. Despite small population sizes, low genetic variability, and no immigration from other streams, these populations have persisted for centuries. A few of the still open questions are: (i) what are the mechanisms allowing the persistence of populations affected by repeated extreme events, and (ii) what are the effects of repeated extreme events on the evolution of life-histories, genetic variability, and phenotypic plasticity in the affected populations?
Adaptive genetic divergence among populations experiencing different stressors occurs rapidly, over small geographical scales even with gene flow, and for traits that directly affect the size and probability of persistence of populations. Rapid genetic and life-history evolution in response to changing trends in climate variables and anthropogenic factors (such as overharvesting, habitat degradation, and habitat fragmentation) is well documented for a variety of species and traits. Adaptive genetic shifts in life-history traits can also be associated with extreme events (e.g., storms, floods, habitat fragmentation or distruction, diseases, biological invasions), but variations in disturbance timing, predictability, frequency, and strength make it difficult to predict the direction and strength of natural selection.
The advent of next-generation sequencing (NGS) technologies made it possible to collect DNA sequence data quickly and at low cost, thus opening a new era of investigation into genome diversity. Variation in genetic markers located in or near functional gene regions can result in changes in the amino-acid sequence of a protein or in its expression, which may affect life-history traits such as mortality, body growth rate, and age at sexual maturity. Investigation of such variation can potentially demonstrate associations between genetic variability, fitness, and environmental factors, as has been successfully shown for some fish species, for example in Atlantic salmon Salmo salar and in Alaskan sockeye salmon Oncorhynchus nerka. In the case of species with expected low levels of genetic variation (e.g., small population sizes), SNPs are better markers than microsatellites, due to the higher abundance of SNPs in the genome. In addition, due to easy standardization and low genotyping error, SNPs are great markers for long‐term, multi-generational parentage studies reconstructing pedigrees in wild populations for evolutionary studies, and they have become the molecular marker of choice in the majority of modern studies on selection at the genome level.
Carry-over effects (COEs) occur when processes in one season influence the some component of an individual's fitness in the following season(s). A particular subset of COEs is the development of phenotypes in response to the environment that they experienced early in life (i.e., in fish between the egg stage and the first stages of independent life). A number of studies have demonstrated trade-offs between early life conditions and fitness in adulthood, but the development of testable mechanistic models has been limited by the complexity of these phenomena, since the interaction between multiple drivers and factors may alter the trajectory of COEs. Early availability of resources seems to induce a plastic response in the body growth of organisms, which become relatively fixed later in life.
Understanding how COEs influence the fitness of individuals and how they interact with density‑dependent processes will strongly contribute to our understanding of population dynamics in animal systems, with implications for conservation and management.
To ameliorate the risks to ecosystems posed by rapid anthropogenic changes in climatic conditions and land use, we need accurate predictions of how species will respond to environmental changes. Correlative models implicitly incorporate biological processes by statistically estimating environment-range associations from occurrence data, while mechanistic models explicitly include biological and ecological processes starting from the vital traits and life histories of the organisms. The integration between correlative and mechanistic models provides a promising “middle ground” approach that could provide more accurate predictions of occupation dynamics in changing environments.
My research is focused on the development of a dynamic framework that integrates correlative and mechanistic species distribution model to assess how the relationships and predictions developed by correlative models can be used in mechanistic models.