Search for: All records

Biogeographic history can set initial conditions for vegetation community assemblages that determine their climate responses at broad extents that land surface models attempt to forecast. Numerous studies have indicated that evolutionarily conserved biochemical, structural, and other functional attributes of plant species are captured in visible-to-short wavelength infrared, 400 to 2,500 nm, reflectance properties of vegetation. Here, we present a remotely sensed phylogenetic clustering and an evolutionary framework to accommodate spectra, distributions, and traits. Spectral properties evolutionarily conserved in plants provide the opportunity to spatially aggregate species into lineages (interpreted as “lineage functional types” or LFT) with improved classification accuracy. In this study, we use Airborne Visible/Infrared Imaging Spectrometer data from the 2013 Hyperspectral Infrared Imager campaign over the southern Sierra Nevada, California flight box, to investigate the potential for incorporating evolutionary thinking into landcover classification. We link the airborne hyperspectral data with vegetation plot data from 1372 surveys and a phylogeny representing 1,572 species. Despite temporal and spatial differences in our training data, we classified plant lineages with moderate reliability (Kappa = 0.76) and overall classification accuracy of 80.9%. We present an assessment of classification error and detail study limitations to facilitate future LFT development. This work demonstrates that lineage-based methods may be a promising way to leverage the new-generation high-resolution and high return-interval hyperspectral data planned for the forthcoming satellite missions with sparsely sampled existing ground-based ecological data. 

Abstract Evolutionary relatedness underlies patterns of functional diversity in the natural world. Hyperspectral remote sensing has the potential to detect these patterns in plants through inherited patterns of leaf reflectance spectra. We collected leaf reflectance data across the California flora from plants grown in a common garden. Regions of the reflectance spectra vary in the depth and strength of phylogenetic signal. We also show that these differences are much greater than variation due to the geographic origin of the plant. At the phylogenetic extent of the California flora, spectral variation explained by the combination of ecotypic variation (divergent evolution) and convergent evolution of disparate lineages was minimal (3%–7%) but statistically significant. Interestingly, at the extent of a single genus (Arctostaphylos) no unique variation could be attributed to geographic origin. However, up to 18% of the spectral variation amongArctostaphylosindividuals was shared between phylogeny and intraspecific variation stemming from ecotypic differences (i.e., geographic origin). Future studies could conduct more structured experiments (e.g., transplants or observations along environmental gradients) to disentangle these sources of variation and include other intraspecific variation (e.g., plasticity). We constrain broad‐scale spectral variability due to ecotypic sources, providing further support for the idea that phylogenetic clusters of species might be detectable through remote sensing. Phylogenetic clusters could represent a valuable dimension of biodiversity monitoring and detection. 

Numerical models based on physics represent the state of the art in Earth system modeling and comprise our best tools for generating insights and predictions. Despite rapid growth in computational power, the perceived need for higher model resolutions overwhelms the latest generation computers, reducing the ability of modelers to generate simulations for understanding parameter sensitivities and characterizing variability and uncertainty. Thus, surrogate models are often developed to capture the essential attributes of the full-blown numerical models. Recent successes of machine learning methods, especially deep learning (DL), across many disciplines offer the possibility that complex nonlinear connectionist representations may be able to capture the underlying complex structures and nonlinear processes in Earth systems. A difficult test for DL-based emulation, which refers to function approximation of numerical models, is to understand whether they can be comparable to traditional forms of surrogate models in terms of computational efficiency while simultaneously reproducing model results in a credible manner. A DL emulation that passes this test may be expected to perform even better than simple models with respect to capturing complex processes and spatiotemporal dependencies. Here, we examine, with a case study in satellite-based remote sensing, the hypothesis that DL approaches can credibly represent the simulations from a surrogate model with comparable computational efficiency. Our results are encouraging in that the DL emulation reproduces the results with acceptable accuracy and often even faster performance. We discuss the broader implications of our results in light of the pace of improvements in high-performance implementations of DL and the growing desire for higher resolution simulations in the Earth sciences. 

Satellite observations show widespread increasing trends of leaf area index (LAI), known as the Earth greening. However, the biophysical impacts of this greening on land surface temperature (LST) remain unclear. Here, we quantify the biophysical impacts of Earth greening on LST from 2000 to 2014 and disentangle the contributions of different factors using a physically based attribution model. We find that 93% of the global vegetated area shows negative sensitivity of LST to LAI increase at the annual scale, especially for semiarid woody vegetation. Further considering the LAI trends ( P ≤ 0.1), 30% of the global vegetated area is cooled by these trends and 5% is warmed. Aerodynamic resistance is the dominant factor in controlling Earth greening’s biophysical impacts: The increase in LAI produces a decrease in aerodynamic resistance, thereby favoring increased turbulent heat transfer between the land and the atmosphere, especially latent heat flux. 

Deep Learning (DL) methods have been transforming computer vision with innovative adaptations to other domains including climate change. For DL to pervade Science and Engineering (S&EE) applications where risk management is a core component, well-characterized uncertainty estimates must accompany predictions. However, S&E observations and model-simulations often follow heavily skewed distributions and are not well modeled with DL approaches, since they usually optimize a Gaussian, or Euclidean, likelihood loss. Recent developments in Bayesian Deep Learning (BDL), which attempts to capture uncertainties from noisy observations, aleatoric, and from unknown model parameters, epistemic, provide us a foundation. Here we present a discrete-continuous BDL model with Gaussian and lognormal likelihoods for uncertainty quantification (UQ). We demonstrate the approach by developing UQ estimates on “DeepSD’‘, a super-resolution based DL model for Statistical Downscaling (SD) in climate applied to precipitation, which follows an extremely skewed distribution. We find that the discrete-continuous models outperform a basic Gaussian distribution in terms of predictive accuracy and uncertainty calibration. Furthermore, we find that the lognormal distribution, which can handle skewed distributions, produces quality uncertainty estimates at the extremes. Such results may be important across S&E, as well as other domains such as finance and economics, where extremes are often of significant interest. Furthermore, to our knowledge, this is the first UQ model in SD where both aleatoric and epistemic uncertainties are characterized. 

The impacts of climate change are felt by most critical systems, such as infrastructure, ecological systems, and power-plants. However, contemporary Earth System Models (ESM) are run at spatial resolutions too coarse for assessing effects this localized. Local scale projections can be obtained using statistical downscaling, a technique which uses historical climate observations to learn a low-resolution to high-resolution mapping. The spatio-temporal nature of the climate system motivates the adaptation of super-resolution image processing techniques to statistical downscaling. In our work, we present DeepSD, a generalized stacked super resolution convolutional neural network (SRCNN) framework with multi-scale input channels for statistical downscaling of climate variables. A comparison of DeepSD to four state-of-the-art methods downscaling daily precipitation from 1 degree (~100km) to 1/8 degrees (~12.5km) over the Continental United States. Furthermore, a framework using the NASA Earth Exchange (NEX) platform is discussed for downscaling more than 20 ESM models with multiple emission scenarios.