Research

Cascade “Ecohydromics” in the Amazonian Headwater System

The Amazon Basin cycles more water through streamflow and evaporation than any other contiguous forest in the world, and transpiration by trees – water taken up by roots and released to the atmosphere – is a critical part of this cycle. Understanding how plant roots, stems, and leaves interact with soil water to jointly regulate forest transpiration across landscapes is a critical knowledge gap, especially as climate changes. Forests are likely adapted to distinct soil moisture conditions in different parts of Amazonian landscapes, such as valleys versus plateaus. To understand landscape hydrology, rainforest species compositions, and their susceptibility to global change, this project studies how water flows are regulated from upstream-to-downstream by plants and soil. This understanding is also critical for Earth-system modeling used to project the fate of Amazonian rainforests and quantify their future influence on climate. This project links diverse disciplines – plant physiology, ecology, hydrology – and integrates them into a model of landscape function. Within this international and interdisciplinary project, my lab focuses on Amazon forest tree physiology in plateaus and valleys.

Specifically, this project will test the hypotheses that: (H1) strong landscape variation in forest transpiration capacity arises from distinct characteristics of trees residing on plateaus (no root access to groundwater) and valleys (root access to groundwater) zones; (H2) previously unquantified “hybrid” soil hydraulics govern soil water fluxes and transit times connecting plateaus and valleys; and (H3) plateau forests influence the composition and function of valley forests by regulating subsurface water flows from higher to lower landscape areas. Study sites are located in the Brazilian Amazon: “KM34” near Manaus, and “KM67” near Santarém. Both sites contain eddy flux towers, canopy access walkways, and a rich history of ecological research and available datasets. Investigators and collaborators include scientists from the University of Michigan, University of Arizona, Universidade Federal do Amazonas, Empresa Brasileira de Pesquisa Agropecuária, Universidade Federal do Oeste Do Pará and Instituto Nacional de Pesquisas da Amazônia among others. This collaborative research project is funded by the National Science Foundation (award number 2106540).

Quantifying leaf-to-landscape predictors of tropical forest drought vulnerability with observations from the International Space Station and mathematical models

Ever since Earth system models first began including carbon-climate interactions, they showed the possibility that tropical forests are vulnerable to changing climate. Dieback of tropical forests would accelerate the increase in atmospheric CO2, whereas resilience of tropical forests could buffer anthropogenic emissions of CO2. Although gaps in theory remain, the greatest challenge to improving model representation of tropical forests may be empirical. Our understanding of physical and biological processes in tropical forests is limited by our ability to test models against observations of tropical forest tree physiology or fluxes of CO2, water, and energy between the canopy and the atmosphere in these diverse systems. These observations must also be informative across spatial scales from ecosystems to continents.

Remote sensing technologies directly relevant to quantifying forest structure, canopy properties, and physiology have been deployed on the international space station (ISS), including GEDI, DESIS, OCO-3, and ECOSTRESS instruments. Our first objective is to integrate observations from these instruments with a biophysical model (SCOPE) to quantify patterns of forest structure, traits, and fluxes (carbon and water) along tropical wet/dry gradients. We will target six tropical forest sites with eddy covariance towers in the Neotropics and Southeast Asia. We will use the DESIS reflectance spectra together with the canopy radiative transfer module of SCOPE to derive a suite of biophysical parameters and vegetation traits. We will validate this approach by comparing output with ground-based measurements of leaf physiology and spectra (at four core sites), and evaluate SCOPE products against fluxes derived from eddy covariance data (all six sites) and benchmarking with artificial neural networks.

Our four core forest sites are 25 to 50 hectare inventory plots (in the ForestGEO network). Our second objective is to use ISS-informed SCOPE to simulate ecosystem water-use efficiency and light use efficiency pixel-by-pixel across these plots to understand variation in driver-flux relationships. The large ForestGEO plot sizes provide continuous information about how the forest changes across topographic gradients, which are in turn associated with shifts in hydrology, soil features, and other physical properties of the plot, many of which reflect larger gradients evident across these biomes. We will test the hypothesizes that (1) low-lying areas with greater potential access to water will show lower ecosystem water-use efficiency, and (2) that light-use efficiency is higher during the dry season in evergreen forests. We will test for the effects of wet / dry season on light-use efficiency using ISS observations, SCOPE, and ForestGEO plot data.

Our third objective will leverage our flux simulations from ISS and SCOPE (from Objective 1) and process understanding (from Objective 2) to extrapolate across regional wet/dry gradients in the Neotropics and Southeast Asian tropics using Harmonized Landsat Sentinel (HLS) data. We will generate look up tables using SCOPE to derive flux estimates along with associated spectral reflectance convolved to match HLS spectral bands.

This project will develop an approach to deriving carbon and water flux estimates through the integration of multiple remote sensing datasets and biophysical modeling in anticipation of future NASA missions that target surface biology with multiple sensors. In addition, the data products (vegetation properties and flux estimates from integration of ISS and SCOPE), and knowledge we gain from them, will contribute to advancing the representation of biophysical processes in forests. Project team members have expertise in tropical forest ecology, ecophysiology, remote sensing, atmospheric physics, and land surface modeling. This project is funded by the NASA Earth Science Division Research and Analysis Program.

Advancing imaging spectroscopy as a tool for organism-to-ecosystem scale forest ecology

My lab is interested in remote sensing tools that can be operated at the tree-to-ecosystem scale from drones and towers to study photosynthesis and carbon cycling. Historically, photosynthesis at scales between individual leaves and expansive forest canopies has been difficult to quantify, with implications for our ability to predict forest responses to climate change. An ideal tool would rapidly collect physiological information at the leaf-to-crown scales and also operate across large areas for studies of how diverse forest communities contribute to carbon and water cycling. Near-surface imaging spectroscopy promises such a tool in the form of solar-induced chlorophyll fluorescence (SIF).

In general, chlorophyll fluorescence is a de-excitation pathway for excited electrons of chlorophyll a that competes with—and thus can offer information about—two other alternative pathways: photochemistry and heat dissipation. SIF is chlorophyll fluorescence when sunlight is the source of excitation energy. Near-surface ‘imaging’ of SIF includes both spatial and spectral information, and from drone or tower platforms can capture the scale of individual tree crowns (e.g. pixels < 0.3 m). SIF must be distinguished from reflected light, and SIF is a small signal (generally <5% of reflected radiance in the near-infrared). I am currently collaborating with the Kellner lab at Brown University and scientists with expertise in optics and calibration to evaluate the interplay of signal, noise, and spatial scale for high resolution SIF imaging spectrometers.