## Illinois Climate Seminar, Spring 2019

The Illinois Climate Seminar aims to bring together researchers in mathematics, statistics, and atmospheric sciences in the region who share common interests in global climate change and regional impacts.

Please contact the organizers, Vera Mikyoung Hur (Mathematics), Bo Li (Statistics), and Ryan Sriver (Atmospheric Sciences) for questions.

All seminars are Wednesdays 3-4pm in 243 Altgeld Hall, unless otherwise indicated.

##### February 6

**Ryan Sriver** (University of Illinois, Atmospheric Sciences)

**Global climate change, regional climate impacts, and quantifying relevant uncertainties**

Earth is warming, and the damages associated with climate and weather extremes (droughts, heatwaves, hurricanes) are increasing. Projecting these changes into the future is difficult due to: incomplete understanding of the physical processes, inadequate numerical models and resolution, and relatively short observational records. Here we highlight some of the current grand climate change problems, and we present some of our group's recent work in areas related to climate extremes and uncertainty quantification surrounding projections of future climate change.

##### February 13

**Chen Chen** (University of Chicago, Geophysical Sciences)

**Warming-induced changes in precipitation organization**

Global warming is well-understood to drive more intense precipitation, but its potential effects on precipitation organization are largely unknown. While some studies suggest that future convective events become larger, others suggest they become smaller. These scenarios imply very different hydrological impacts. Using an object-based approach, we show that in high-resolution models, warming forces summer rainstorm profiles to intensify but concentrate across the United States. We demonstrate that the organization changes can be proxied by changes in rainstorm intensity and size given threshold-based rainstorm boundaries, and the disagreement in the literature is likely an artifact of different threshold choices. The tendency to stronger but smaller rainstorms manifests across all sizes of convective events and results in a robust bivariate distributional shift. It is analogous to the seasonal variation, in which rainstorms become stronger but smaller as summer warms. These organization changes reflect temperature-driven responses of convective precipitation, which are clearly distinguishable from the natural variation when stronger rainstorms are often in larger sizes.

##### February 20

**Elisabeth Moyer** (University of Chicago, Geophysical Sciences)

**Climate science in the age of big data**

##### February 27

**Sooin Yun** (University of Illinois, Statistics)

**Detection of local differences between two spatiotemporal random fields**

Comparing the characteristics of spatiotemporal random fields is often at demand. However,
the comparison can be challenging due to the high-dimensional feature and dependency in the
data. We develop a new multiple testing approach to detect the local difference in the
characteristics of two spatiotemporal random fields by taking the spatial information into
account. Our method adopts a two-component mixture model for location wise p-values and
then derives a new false discovery rate (FDR) control, called mirror procedure, to determine the
optimal rejection region. This procedure is robust to model misspecification and allows for
weak dependency among hypotheses. To integrate the spatial heterogeneity, we model the
mixture probability as well as allow the alternative distribution to be spatially varying. An EM-
algorithm is developed to estimate the mixture model and implement the FDR procedure. We
study the FDR control and the power of our new approach both theoretically and numerically,
and finally apply the approach to compare the mean and teleconnection pattern between two
synthetic climate fields. This is a joint work with Xianyang Zhang and Bo Li.

##### March 6

**Ben Vega Westhoff** (University of Illinois, Atmospheric Sciences)

**Application of the Hector simple climate model for probabilistic assessment of temperature and sea-level rise**

Simple climate models are useful tools for quantifying decision-relevant uncertainties, given their flexibility, computational efficiency, and suitability for large-ensemble frameworks necessary for statistical estimation using resampling techniques (e.g. Markov chain Monte Carlo?MCMC). First, I give a brief summary of selected past applications of simple climate models. Then I summarize our current project using the simple, open-source, global climate model Hector, coupled with a sea-level change module (Building blocks for Relevant Ice and Climate Knowledge; BRICK) that also represents contributions from thermal expansion, glaciers and ice caps, and polar ice sheets. We apply a Bayesian calibration approach to quantify model uncertainties surrounding 39 model parameters, using observational information from global surface temperature, ocean heat uptake, and sea-level change. The calibrated model is then used to perform probabilistic projections of global temperature and sea-level rise.

##### March 13

**Trevor Harris** (University of Illinois, Statistics)

**Testing the exhangeability of two spatiotemporal processes with applications to evaluating proxy influence in data assimilation**

Statistical inference on spatiotemporal processes is a fundamental problem in many fields including Ecology, Oceanography, and Climatology. Of particular interest to the paleoclimate community is the study of Climate Field Reconstructions (CFRs) with seasonal to annual resolution spanning the last several millennia. CFRs attempt to recover spatiotemporal fields of climate variables, using proxy records of past climate variability, and have emerged as important tools for studying the mechanisms of climate change. Motivated by assessing differences between CFRs, we propose a new method for evaluating the differences in the distributions of two spatiotemporal processes by using the notions of data depth and functional data. Our test is robust, computationally efficient, distribution free and has a convenient asymptotic distribution. We apply our test to study global and regional proxy influence on a Data Assimilation based CFR by comparing its background and analysis states. We find that there is a steadily increasing divergence between the state?s distributions over time, indicating increasing proxy influence, and that proxy influence can extend far beyond collection sites.

##### March 27

**Matthew Huber** (Purdue University, Earth, Atmospheric and Planetary Sciences)

**Life in a steambath**

Heat stress arises from a combination of temperature, humidity, and radiative load. I describe a framework for thinking about heat stress in a warmer world and place that within the context of what we know about the dynamic variability of climate on long time scales. Moist heat stress is tightly constrained by atmospheric dynamics and thermodynamics, but radiative load is not so directly constrained. Regional scale applications will be shown both for the Northeast U.S. and India.

##### April 3

**Alfredo Wetzel** (University of Wisconsin-Madison, Mathematics)

**Discontinuous fronts as exact solutions to precipitating quasi-geostrophy**

Atmospheric fronts may be idealized as boundaries between two air masses with different temperature, density, moisture, etc. In this presentation, we discuss exact discontinuous solutions of a simplified model for moist mid-latitude synoptic atmospheric flows, the precipitating quasi-geostrophic (PQG) equations. These simple discontinuous solutions correspond to propagating moist fronts that require both rainfall and a phase change of water at the front interface to exist. The fronts propagate at speeds related to the rainfall velocity, temperature/wind jump magnitudes, and front geometry. Moreover, the relative simplicity of the model and front geometry gives rise to readily accessible conditions for the front?s existence. As an initial assessment of the realism of these fronts, we use rough estimates of relevant physical parameters to show that cold, warm, and stationary fronts are captured by the model. Simple exact solutions have previously been presented in the context of the well-known Margules? front slope formula for dry fronts but have not been generalized to include propagation and moisture.

##### April 10

**Danielle Sass** (University of Illinois, Statistics)

**Return level estimation for large spatial extremes**

An important goal of extremes modeling is to estimate the T-year return level together with its confidence interval (CI). Spatial extremes are common for climate data as the observations are usually referenced by geographic locations and correlated when they are nearby. The return level estimation for spatial extremes relies on models particular suitable for spatial extremes including max-stable models, copula, Bayesian methods, and spatial generalized extreme value (GEV). Among those methods, spatial GEV is the simplest and fastest approach at the price of ignoring the correlation among observations, yet simulations show that return level estimation using spatial GEV still provides satisfactory results compared to other computationally intensive methods. However, the usual assumption for shape parameter to be an unknown constant over the whole spatial domain may hamper its application to large spatial extremes as this assumption likely becomes unrealistic while the return level estimation is sensitive to the shape parameter. We propose a fast approach based on spatial GEV that allows the shape parameter to vary smoothly over the spatial domain using fused lasso and fused ridge. Bootstrap methods are applied to estimate the CI. This provides a feasible way to estimate return levels for large spatial extremes.

##### Tuesday, April 23, 4-5pm in 245 Altgeld Hall -- MSS Colloquium

(Note date, time, and location.)

**Mary Silber** (University of Chicago, Statistics)

**The challenge of modeling dryland vegetation pattern formation using ideas from dynamical systems**

A beautiful example of spontaneous pattern formation appears in the distribution of vegetation in some dry-land environments. Examples from Africa, Australia and the Americas reveal that vegetation, at a community scale, may spontaneously form into stripe-like bands, alternating with striking regularity with bands of bare soil, in response to aridity stress. A typical length scale for such patterns is 100 m; they are readily surveyed by modern satellites (and explored from your armchair in Google maps). These ecosystems represent some of Earth?s most vulnerable under threats to desertification, and some ecologists have suggested that the patterns, so easily monitored by satellites, may have potential as early warning signs of ecosystem collapse. I will describe efforts based in simple mathematical models, inspired by decades of physics research on pattern formation, to understand the morphology of the patterns, focusing particularly on topographic influences. I will take a critical look at the role of mathematical models in developing potential remote probes of these ecosystems. How does mathematical modeling influence what we see? Does it suggest what we should monitor? Could it lead us astray?

##### April 24

**Mary Silber** (University of Chicago, Statistics)

**Pattern formation in the drylands: Vegetation patterns in mathematical models and in satellite images**

Abstract: An awe-inspiring example of spontaneous pattern formation appears in the distribution of vegetation in some dry-land environments. Examples from Africa, Australia and the Americas reveal vegetation congregated in stripe-like bands, alternating with bands of bare soil. A typical length scale for such patterns is ~100 m; they may be readily surveyed in Google Maps. The typical time scale for pattern evolution, however, is ~100 years, so investigations of dynamics are a bit thwarted, with only a few early data points provided by aerial photographs from the 1950s. This talk will highlight investigations of image data, focusing in particular on our studies of vegetation patterns in the Horn of Africa. The goal of this presentation is to highlight the potential for models to be confronted by data, and vice versa.