Hybridization is thought to have played an important role in shaping the evolutionary history of diverse island taxa. Yet secondary contact doesn’t always result in introgressive hybridization, and it is important to understand what determines the ecological and evolutionary outcomes of secondary contact, particularly in the face of widespread secondary contact between island endemics and recently introduced species. While there are reasons that secondary contact commonly leads to heterospecific mating on islands, the consequences of secondary contact will differ depending on the nature of the species involved. I will present the results of our field work on introduced and endemic species of the plant Psidium (guava) on the Galápagos Islands of Ecuador, exploring the role of secondary contact in the ongoing extirpation of the endemic species. I will then discuss the results from a quantitative analysis of published empirical research on secondary contact among vertebrate, invertebrate, and plant species on the five most well-studied remote oceanic archipelagoes. This analysis suggests the relative importance of different factors driving secondary contact and heterospecific mating, such as disturbance, inter-island dispersal, and compromised assortative mating. I will discuss the hypothesis that introgression is a more common outcome between island endemic species while reproductive interference is a more common outcome between endemic and introduced species, as well as the contention that reproductive interference between endemic and introduced species is a cryptic threat to the conservation of island endemics.

The terrestrial carbon (C) cycle involves fluxes between multiple pools that determine ecosystem functions and regulate global climate. These fluxes and pools are influenced by changes in abiotic (temperature, precipitation etc.) and biotic (animals, microbes, etc.) factors. In this thesis, I address three questions on how these abiotic and biotic drivers influence the size and stability of these fluxes and pools.    

            In the first chapter, I investigate how covariation between decadal trends (2001-2019) in temperature and precipitation influence the two opposing C-fluxes in the soil-C pool – (1) C-influx through primary production (NPP), and (2) C-efflux through soil heterotrophic respiration (Rh). I estimate how any imbalance between these opposing fluxes affects the vulnerability of soil-C across the globe. I find that changes in C-influx may not compensate for rising C-efflux, under wetter and warmer conditions. Soil-C loss can occur in both tropics and at high latitudes, and precipitation emerged as the key determinant of soil-C vulnerability in a warmer world. This implies that hotspots for soil-C loss/gain can shift in the coming decades to make the soil-C pool vulnerable to climate change despite widespread increase in NPP across the world. 

            In the second chapter, I explored the influence of climate on long-term correlations in vegetation fluctuations (i.e., persistence, measured by the Hurst exponent from time-series data). I found evidence for stronger persistence in warm and dry regions of the world, and there were non-linear relationships between persistence and two key climate variables (i.e., temperature and precipitation). While average temperature and precipitation together explained nearly three-fourths of the spatial variation in vegetation persistence, they had limited ability to explain the observed temporal changes in persistence. We find evidence for change in vegetation persistence across the globe driven by background change in climate. This provides some new insights into the resistance/resilience of vegetation in different ecosystems.  

In the third chapter, I investigated how animals – large mammalian herbivores – influence the terrestrial carbon cycle. Their influence on the size of the soil-C pool is well known, but how herbivores control the temporal stability of soil-C has remained largely unknown. I used a long-term field experiment in the Trans-Himalaya (2005-present) to estimate the consequences of herbivore-exclusion on interannual fluctuations in soil-C. I found high interannual variability in soil-C, and herbivores promote temporal stability of soil-C. Grazing by herbivores also mediated the influence of nitrogen on the stability on soil-C. Therefore, conserving large mammalian herbivores in grazing ecosystems can help achieve nature-based climate solutions. 

            Overall, this thesis explores the linkages between three aspects of the terrestrial carbon cycle and climate, and herbivores. It highlights the non-linearities in vegetation-soil-animal interactions which are important for the stability of the terrestrial carbon cycle.

Soil contains 2500 Petagrams of Carbon, which is two to three times as much as the Carbon in the atmosphere, or in land vegetation. The soil carbon pool receives organic matter from the land vegetation pool in the form of leaf litter, dead organisms and animal excreta. In turn, it loses carbon dioxide to the atmosphere through microbial respiration. Since the Industrial Revolution, atmospheric carbon dioxide levels have risen by 50% (from 280 to 421 parts per million), and global temperature has risen by about 1.1 degrees. Further, there is concern that rising temperatures may increase soil microbial respiration and hasten the release of carbon to the atmosphere, creating a positive feedback loop. We therefore need to study the rate at which carbon is lost from soil, and how human activities - such as land use patterns and agricultural practices - may alter it. In this thesis, I propose to study how much carbon is present, and how much is respired, from soil in the Spiti valley region of Himachal Pradesh, which is a part of the carbon-rich Trans-Himalayan rangelands. I will also look at how land use, fertilisation and warming affect soil respiration. The first chapter of this thesis will look at what climatic, ecological and anthropogenic factors influence soil respiration. Since soil respiration in different environments depends also on the substrates available, I will review published literature on catabolic response profiling, which measures soil respiration against various metabolic substrates. The second chapter will focus on understanding the spatial and chemical structure of soil pools and fluxes in Spiti. Current soil Carbon models, constructed based on principles of reaction kinetics, do not account for spatial heterogeneity. We plan to measure the amount of Carbon in various soil pools, and to perform lab incubations at different temperatures to understand the sensitivity of their fluxes to temperature. We will also distinguish between the two main soil carbon pools - mineral-associated and particulate organic matter. These pools have different flux rates and spatial distributions, and so would respond differently to changes in land use or temperature. The final chapter will look at how land use affects these various carbon pools. Land grazed by native wildlife in Spiti has been converted to livestock pasture and agricultural fields, and the global fertilisation experiment NutNet maintains plots with various types of fertiliser application. Differences in the magnitude and flux of each carbon pool will be measured for each treatment. This thesis is expected to improve our current understanding of the soil carbon loss by respiration, and to predict how land use change and fertilisation will influence the rate of this loss. This information will be useful in modelling climate change and soil quality, as well as in the making of policies related to land conversion and agricultural practices.

Marine capture-fisheries have expanded ten-fold over the past few decades leading to severe decline in fish stocks as well as populations of non-target or ‘bycatch’ species, which comprise ~ 40% of global catch. Large-bodied, slow growing species, like elasmobranchs, cetaceans, and sea turtles, referred to as marine megafauna, which generally occupy higher trophic levels, are perhaps most severely affected by such mortality. At least 160 species of elasmobranchs and 25 species of cetaceans have been recorded from Indian waters, several of which are regularly caught in fisheries. However, information on them is limited to their catch rates with little data on the drivers of their bycatch. Incidental mortality of cetaceans is even more challenging to record since all species are protected by law and bycatch is discarded at sea.  For my research, I propose to understand how the spatial and temporal patterns of fisheries impact marine megafauna in Indian waters, and to what extent may these be explained by ecological and anthropogenic factors. The chapters of my thesis focus on a) Combining fish-landing site surveys with spatial information on multi-gear fisheries to understand bycatch risk of elasmobranchs at two contrasting locations i.e., Malvan and Visakhapatnam on the west and east coast of India, respectively to test whether bycatch patterns can be generalised across environments, b) testing novel methods to combine oceanographic particle tracking with hierarchical modelling to quantify and elucidate the broader drivers of cetacean bycatch using ocean circulation, remote sensing, and stranding data across India; and c) understanding the behaviour of Indian Ocean humpback dolphins around fishing vessels to explore specific mechanisms of bycatch in Goa, potentially using drone based surveys.   For my first chapter, I have sampled more than 1500 fishing vessels across two seasons at both sites. Gear use and diversity are similar across the sites, but species composition and catch-rates differ. For my second chapter, preliminary results using dead cetacean stranding data from Goa suggest that between 2 - 4 small cetaceans may be dying off its coast per 100 km2 per year, of which at least 30% may be due to bycatch in fishing gear. Altogether, the results of my thesis will elaborate on how and why megafauna species are susceptible to bycatch and help develop plans for species conservation and fisheries management. 

Passive Acoustic Monitoring (PAM) has emerged as a possible alternative to traditional field surveys, which are tedious, expensive, and often limited to a few sites. Although PAM eases collection of data, the sheer volume of data generated complicates the analysis process. Recently, researchers have simplified the analysis process by using soundscape-based machine learning models that can extract information such as species richness and occurrence from audio data. But most of these methods either fail to generalize to various landscapes and across anthropogenic gradients or have not been tested in such environments. Thus, for my first objective, I will develop soundscape-based methods to determine habitat quality across different tropical landscapes. I aim to collect audio data across different tropical habitats along a gradient of land use and test the performance of current soundscape-based models and build better generalizable models. Soundscape-based methods have also been able to track spatiotemporal avian biodiversity patterns in temperate climates at coarser timescales and infer species occurrence. Thus, for my second objective, I will test if we can use these methods to track avian spatiotemporal biodiversity patterns at finer timescales in tropical habitats, and if we can use soundscapes to predict presence of non-vocalizing species. I aim to collect fine scale audio data along with ground truthed bird species observations and test performance of these algorithms on the data. I also aim to build a model that can use soundscapes to predict the occurrence of non-vocalizing species. Studies have shown that insects are the most vocalizing species in tropical habitats, however, most studies don't focus on insect vocalizations. This can be attributed to the fact that building automated models for insect species identification using acoustic data is a challenging problem. Thus, for my third objective, I will use recent developments in the field of machine learning, such as transformers, to build acoustic identification models for insects and quantify insect biodiversity.

Coastal mangrove ecosystems remain among Earth’s most valuable forests, serving as key carbon sinks and protecting billions of dollars of coastal infrastructure from sea level rise. Mangroves are also highly vulnerable to both direct human influence and growing climatic threats, with hotspots of loss scattered across the tropics. This seminar will present the groundbreaking methodology and results from a global-scale satellite-based analysis to map the drivers and extent of mangrove losses over the last several decades. Using a machine learning analysis of over one million Landsat images, we capture mangrove losses due to agriculture, urbanization, sea level rise, and extreme weather events at high resolution. This analysis is currently being integrally used to monitor global restoration and conservation initiatives, inform key mangrove preservation policies, and evaluate progress on the Sustainable Development Goals.

This seminar will also provide participants with an overview of a large-scale capacity building initiative to train thousands of students across South Asia and West Africa with cutting-edge remote sensing tools for environmental monitoring. Participants will have the opportunity to directly get involved in training and collaboration opportunities with NASA, National Geographic, Google, and Stanford University in the field of satellite-based climate analysis.

Temperatures are rising and the thermal regimes of our ecosystems are changing rapidly, becoming warmer and more variable with more frequent extreme events. The impacts of these changes are the subject of much current research on a wide variety of organisms. Important questions revolve around what determines thermal tolerance or the thermal limits of a species distribution. Are they the same thing? And how best to measure or predict them? To assess the integrated response of organisms and communities to changing thermal regimes requires systems level perspectives that draw together impacts at different scales – e.g., cell vs organ vs organism within species; or plant vs pollinator vs pathogen at community scales. I will present work collaborators and I have been doing to better understand what thermal tolerance means, what drives it, and to explore how best to assess it.

Over 80 % of the world animals have complex life-cycles with multiple life-stages.  In many of these species, each life stage experiences dramatically different selection pressures, due to living in different environments and specialising in different tasks. For example, caterpillars must avoid predation while feeding on leaves, while butterflies must fly, find mates and feed on nectar. However, all butterflies were once caterpillars, and they share the exact same genome. This raises a fundamental evolutionary question – to what extent does the resetting of the body plan at metamorphosis allow for independent adaptation at each life stage? Answering this question is essential to predict how animals might adapt to environmental change. In this talk I will describe a recent collaboration looking at the evolution of colouration across life-stages of Australian Shield bugs, as well as future research plans to investigate the evolution of thermal tolerance in Australian leaf beetles.

Tropical forests are currently undergoing large-scale structural changes, including an increase in liana abundance and biomass. As critical but frequently overlooked components of these forests, lianas can reduce tree growth and increase tree mortality, thus significantly influencing the declining pantropical carbon sink potential. Despite the increasing importance of lianas, quantitative studies on liana abundance and their impact on forest carbon stocks are almost non-existent. Recent advancements in terrestrial laser scanning (TLS) technologies have provided a new lens through which we can examine forest structures. These developments facilitate observation on an unprecedented scale and offer improved accuracy. As a result, we can now reliably measure liana structure and assess its impact on the structure and functionality of tropical forests. In this presentation, I will elaborate on how the utilization of TLS, in conjunction with other recent technological innovations in the remote sensing field, has enhanced our ability to understand lianas. Additionally, I will share the insights we have gained about the role lianas play in contributing to the already declining pantropical carbon sink potential.  

Environments shape the development of physiology but also act as potent selective sieves to shape how physiological responses evolve. I’ll briefly discuss the work my group has been doing to 1) test how early thermal and resource environments shape thermal physiology in a model lizard system (Lampropholis delicata); 2) the generality of such findings across ectotherms and 3) some new work that explores the interplay between physiological plasticity and the opportunity for selection on physiological responses in an era of climate change.