Shyam Thomas
Geospatial Data Manager & Analyst
The Nature Conservancy (Minnesota - North Dakota - South Dakota)
About Me
I am a broadly trained ecologist with a long-standing interest in natural history & conservation. I acquired a BS in Botany & Zoology from St Xavier's college Mumbai (India) and an MS in Plant Ecology from Oklahoma State University under the mentorship of Michael Palmer. After a short stint as Information Officer at the Bombay Natural History Society, I returned to the US to pursue Ph.D. in landscape ecology and modeling in Kirk Moloney's lab at Iowa State University.
After my Ph.D., I was a postdoc in Matt Daugherty's lab at the University of California Riverside. With Matt and collaborators from USDA, I worked on developing spatiotemporal models aimed at understanding the landscape-level dynamics of a recent insect invasion. In my second postdoc with Stephanie Melles at Ryerson University, Toronto I developed geospatial visualization tools and structural equation models to improve on our current understanding of mercury contamination in sport fishes. Currently, I am a postdoc in the labs of Gretchen Hansen and Dan Larkin at the University of Minnesota researching the impact of invasive Eurasian Watermilfoil.
Drivers of invasive Eurasian watermilfoil distribution & abundance
Species distribution models typically involve species occurrence information as the response data, especially in the form of widely available presence-only datasets. Systematically collected information on species presence, absence, and abundance remain a challenge to find. Understandably, SDMs have seldom combined the strengths of presence-absence data together with abundance data. Here we avail a uniformly collected lake survey data on invasive aquatic macrophytes. Focussing on invasive Eurasian watermilfoil, we developed a series of SDMs (using Random Forest algorithm) with three distinct response data types: (i) presence-only data, (ii) presence-absence data, and (iii) abundance data. Our findings highlight interesting differences between occurrence-based and abundance-based SDMs. Specifically, drivers of invasion risk (occurrence probability) and invasion impact (abundance) are not always the same. Moreover, evaluation of presence-only & presence-absence SDMs using abundance data as a measure of functional accuracy underscored the value of absence information and the limitations of conventional discrimination accuracy metrics (such as AUC). For more on this, see our recent research article published in Ecology & Evolution
Completed Research Projects
Mercury in fish: a study of pattern and process across multiple scales
Mercury in fish is a major environmental concern as it is a direct threat to human health and well-being. Organic methylmercury in fish bioaccumulates, such that larger individuals tend to have higher amounts of mercury. Fish mercury levels are thus predicted as a function of body size. Additional factors like climate, surrounding landscape conditions, lake water chemistry, trophic position, species composition can further modulate mercury bioaccumulation. Inter-and within-lake variations are thus expected. However, these variations are not well quantified, and more importantly, the significance of the various multi-scale drivers and their interactions are poorly understood.
At Ryerson University, I collaborated with the Ontario Ministry of Environment & Climate Change to study mercury bioaccumulation in Walleye and Northern Pike. Using mixed-effects models, spatiotemporal variations in mercury bioaccumulation were quantified at fine and broad scales. Fine-scale models at the lake and year level explained most of the variations. Broad-scale models at 3 latitudinal degrees and 5-year time periods provided a useful index that summarized mercury trends in a climate change context. For an improved understanding of the role of various drivers and processes operating across multiple scales (climate-, landscape-, lake-, & fish- levels), structural equation models were used. Climate and landscape conditions were found to impact fish mercury levels, but indirectly by altering lake water chemistry and fish size.
Inferring invasion process from spatiotemporal patterns
Biological invasions are characterized by dynamic spatiotemporal patterns and associated processes that are most evident at the landscape level. However, landscape-level distribution patterns of invasive species are rarely leveraged to comprehensively understand the process of invasion. Here I address the potential of spatiotemporal distribution patterns in understanding the process of invasion by considering a stage-based approach for an invasive insect. In short, the recent invasion of Asian Citrus Psyllid (a destructive insect pest) in southern California and the associated spatiotemporal patterns were analyzed to characterize three key stages of invasion.
As part of this (postdoc) research, I worked with a large monitoring database of ACP collected across urban South California. The extensively collected spatiotemporal data on ACP invasion since 2008 provides a unique opportunity to understand an active ongoing invasion process and characterize key stages of invasion - introduction, establishment, spread, and impact. Using a suite of spatiotemporal analyses and habitat suitability models, the study addresses three key questions: (i) Do major roads serve as long-distance dispersal corridors of ACP introduction in an urban landscape?, (ii) Can habitat suitability models based on presence and abundance information capture ACP's establishment and impact risks, respectively?, and finally (iii) Do the locations predicted as high establishment and high impact risk overlap?
Effects of aboveground herbivory on plants with stored belowground biomass
Herbivory is perhaps the most ubiquitous of all ecological interactions with a long history of empirical and theoretical studies. One aspect of herbivory that continues to puzzle researchers is - how sessile plants defend and sustain repeated (often severe) damage caused by herbivores? 'Tolerance' is one such well-recognized defense mechanism where plants can maintain fitness after damage by regrowth, adaptive allocation of resources, or enhanced reproduction. Allocation of biomass to roots in response to aboveground defoliation damage is considered common among perennials with large perennating rootstocks.
As part of my Ph.D. research, I developed a theoretical model that explored how temperate perennials with strong seasonality in growth and long-term stored biomass tolerate periodic herbivory by specialist insect folivores. The study addresses how three key plant traits - spring growth from the rootstock, post-herbivory regrowth, and eventual allocation of biomass to roots - affect plant tolerance response and folivore population dynamics. Results from the coupled plant-insect population model indicate that allocation to belowground storage is a critical factor in determining a plant's tolerance response, and it also affects folivore population dynamics (Thomas et al. 2017). From an applied perspective, findings suggest that invasive perennials with efficient storage allocation are weakly impacted by defoliating insects, which questions the effectiveness of conventional biocontrol programs.
Species distribution models: hierarchical effects of land-use & propagule pressure
The study of species distribution is one of the oldest and central premises in ecology that strives to explain biodiversity patterns in space and time. Earlier, biogeographical studies considered mostly factors such as climate that operate at large spatial scales. Species distributions are now increasingly viewed as an 'emergent' outcome of multiple factors (biotic and abiotic) operating hierarchically in space. Not surprisingly, species distribution models (aka habitat suitability/niche models) now explicitly incorporate ecological factors in a hierarchical framework, wherein effects of various factors are scale-dependent (see above figure).
My Ph.D. dissertation project explored the spatial distribution of an invasive wetland plant - purple loosestrife (inset picture above) by taking a hierarchical approach. My research highlighted that loosestrife invasion across an urban landscape is a complex process wherein three key spatially nested ecological factors - availability of suitable habitat (i.e. wetlands), its spatial neighborhood, and propagule pressure, together explain loosestrife distribution. Predictive invasive species models based on surrounding land-use conditions and propagule pressure further vindicated the significance of incorporating fine-scale processes such as dispersal, since models that included propagule pressure as the eventual nested factor yielded accurate predictions and also minimized the area predicted as high risk of invasion.
Community ecology & conservation of grassland ecosystems
Grasslands are interesting ecosystems in that they are largely maintained by a suite of natural disturbances like fire & grazing. Not surprisingly, grasslands have provided a 'natural' platform to study and understand how various disturbances, both natural & anthropogenic, modulate ecological systems. Given the increasing trend in anthropogenic modifications of natural habitats and disturbance-driven establishment of exotic species, understanding the 'ecology of disturbances' is today all the more significant. And, grassland systems continue to provide an ideal ecological setting for this purpose.
My masters research project explored how species diversity and composition respond to disturbances in an Oklahoman tallgrass prairie grassland, through a long-term experimental mowing study. The study highlighted the usefulness of mowing as a restoration tool as it enhances species richness by creating microsites or gaps with high light availability. The study also revealed the confounding effects invasive plant species can have on grassland conservation as disturbances that enhance plant diversity can also facilitate the establishment of invasive plants (Dee et al. 2016). Alongside, I also reviewed the literature on the lesser-known montane shola-grasslands of Western Ghats, India in an effort to understand the community ecology and conservation lessons learned and identify key gaps that remain (Thomas & Palmer 2007).
Publications
(click on journal thumbnails for pdf copy)
Schartel, T. E., Bayles, B. R., Cooper, M. L., Simmons, G. S., Thomas, S. M., Varela, L. G., & Daugherty, M. P. 2019. Reconstructing the European Grapevine Moth (Lepidoptera: Tortricidae), Invasion in California: Insights From a Successful Eradication. Annals of the Entomological Society of America, 112(2), 107-117.
Thomas, S.M., Melles, S.J. & Bhavsar, S.P. 2018. Spatiotemporal variations in mercury
bioaccumulation at fine and broad scales for two freshwater sport fishes Water 10, 1625;
https://doi.org/10.3390/w10111625Useful online resources
“The world cannot be understood without numbers. But the world cannot be understood with numbers alone.” ― Hans Rosling (Factfulness)
Always keen to hear from folks with similar interests.