Columbia Water Center Researchers to Present at AGU Annual Fall Conference 2017

Columbia Water Center researchers will be among the hundreds of world-class scientists presenting at the American Geophysical Union’s annual fall conference. The conference will be held from December 11-15 in New Orleans. Please see below for the scheduled events that our team will be organizing or participating in.

_____________________________________________________________________

H11H-1301: Big Data and Heath Impacts of Drinking Water Quality Violation

Maura Allaire, Shuyan Zheng, Upmanu Lall

Monday, 11 December 2017  08:00 – 12:20

New Orleans Ernest N. Morial Convention Center – Poster Hall D-F

 

Health impacts of drinking water quality violations are only understood at a coarse level in the United States. This limits identification of threats to water security in communities across the country. Substantial under-reporting is suspected due to requirements at U.S. public health institutes that water borne illnesses be confirmed by health providers. In the era of ‘big data’, emerging information sources could offer insight into waterborne disease trends.

In this study, we explore the use of fine-resolution sales data for over-the-counter medicine to estimate the health impacts of drinking water quality violations. We also demonstrate how unreported water quality issues can be detected by observing market behavior. We match a panel of supermarket sales data for the U.S. at the weekly level with geocoded violations data from 2006-2015. We estimate the change in anti-diarrheal medicine sale due to drinking water violations using a fixed effects model.

We find that water quality violations have considerable effects on medicine sales. Sales nearly double due to Tier 1 violations, which pose an immediate health risk, and sales increase 15.1 percent due to violations related to microorganisms. Furthermore, our estimate of diarrheal illness cases associated with water quality violations indicates that the Centers for Disease Control and Prevention (CDC) reporting system may only capture about one percent of diarrheal cases due to impaired water. Incorporating medicine sales data could offer national public health institutes a game-changing way to improve monitoring of disease outbreaks. Since many disease cases are not formally diagnosed by health providers, consumption information could provide additional information to remedy under-reporting issues and improve water security in communities across the United States.

_____________________________________________________________________

H11J-1340: Financial Risk Reduction and Management of Water Reservoirs Using Forecasts: A Case for Pernambuco, Brazil

Ipsita Kumar, Laureline Josset, Upmanu Lall

Monday, 11 December 2017  08:00 – 12:20

New Orleans Ernest N. Morial Convention Center – Poster Hall D-F

 

The financial health and sustainability, ensuring adequate supply, and adapting to climate are fundamental challenges faced by water managers. These challenges are worsened in semi-arid regions with socio-economic pressures, seasonal supply of water, and projected increase in intensity and frequency of droughts. Over time, probabilistic rainfall forecasts are improving and for water managers, it could be key in addressing the above challenges. Using forecasts can also help make informed decisions about future infrastructure.

The study proposes a model to minimize cost of water supply (including cost of deficit) given ensemble forecasts. The model can be applied to seasonal to annual ensemble forecasts, to determine the least cost solution. The objective of the model is to evaluate the resiliency and cost associated to supplying water.

A case study is conducted in one of the largest reservoirs (Jucazinho) in Pernambuco state, Brazil, and four other reservoirs, which provide water to nineteen municipalities in the Jucazinho system. The state has been in drought since 2011, and the Jucazinho reservoir, has been empty since January 2017. The importance of climate adaptation along with risk management and financial sustainability are important to the state as it is extremely vulnerable to droughts, and has seasonal streamflow.

The objectives of the case study are first, to check if streamflow forecasts help reduce future supply costs by comparing k-nearest neighbor ensemble forecasts with a fixed release policy. Second, to determine the value of future infrastructure, a new source of supply from Rio São Francisco, considered to mitigate drought conditions.

The study concludes that using forecasts improve the supply and financial sustainability of water, by reducing cost of failure. It also concludes that additional infrastructure can help reduce the risks of failure significantly, but does not guarantee supply during prolonged droughts like the one experienced currently.

_____________________________________________________________________

H12D-02: Enhancing Groundwater Cost Estimation with the Interpolation of Water Tables across the United States

Ain Ul Mardhiah Rosli, Upmanu Lall, Laureline Josset, Tess A Russo, Tristan Eisenhart

Monday, 11 December 2017  10:35 – 10:50

New Orleans Ernest N. Morial Convention Center – 280-282

 

Analyzing the trends in water use and supply across the United States is fundamental to efforts in ensuring water sustainability. As part of this, estimating the costs of producing or obtaining water (water extraction) and the correlation with water use is an important aspect in understanding the underlying trends. This study estimates groundwater costs by interpolating the depth to water level across the US in each county. We use Ordinary and Universal Kriging, accounting for the differences between aquifers. Kriging generates a best linear unbiased estimate at each location and has been widely used to map ground-water surfaces (Alley, 1993).The spatial covariates included in the universal Kriging were land-surface elevation as well as aquifer information. The average water table is computed for each county using block kriging to obtain a national map of groundwater cost, which we compare with survey estimates of depth to the water table performed by the USDA. Groundwater extraction costs were then assumed to be proportional to water table depth. Beyond estimating the water cost, the approach can provide an indication of groundwater-stress by exploring the historical evolution of depth to the water table using time series information between 1960 and 2015. Despite data limitations, we hope to enable a more compelling and meaningful national-level analysis through the quantification of cost and stress for more economically efficient water management.

_____________________________________________________________________

B12D-05: Representing Plant Hydraulics in a Global Model: Updates to the Community Land Model

Daniel Kennedy, Pierre Gentine

Monday, 11 December 2017  11:20 – 11:35

New Orleans Ernest N. Morial Convention Center – 386-387

 

In previous versions, the Community Land Model has used soil moisture to stand in for plant water status, with transpiration and photosynthesis driven directly by soil water potential. This eschews significant literature demonstrating the importance of plant hydraulic traits in the dynamics of water flow through the soil-plant-atmosphere continuum and in the regulation of stomatal aperture. In this study we install a simplified hydraulic framework to represent vegetation water potential and to regulate root water uptake and turbulent fluxes.

Plant hydraulics allow for a more explicit representation of plant water status, which improves the physical basis for many processes represented in CLM. This includes root water uptake and the attenuation of photosynthesis and transpiration with drought. Model description is accompanied by results from a point simulation based at the Caxiuanã flux tower site in Eastern Amazonia, covering a throughfall exclusion experiment from 2001-2003. Including plant hydraulics improves the response to drought forcing compared to previous versions of CLM. Parameter sensitivity is examined at the same site and presented in the context of estimating hydraulic parameters in a global model.

_____________________________________________________________________

B13H-1848: Influence of Drought on the Hydraulic Efficiency and the Hydraulic Safety of the Xylem – Case of a Semi-arid Conifer.

Marceau F Guerin, Kevin Lee Griffin, Laia Andreu-Hayles, Pierre Gentine

Monday, 11 December 2017  13:40 – 18:00

New Orleans Ernest N. Morial Convention Center – Poster Hall D-F

 

Recent droughts in the Southwest US have resulted in extensive mortality in the pinion pine population (Pinus Edulis). An important factor for resiliency is the ability of a plant to maintain a functional continuum between soil and leaves, allowing water’s motion to be sustained or resumed. During droughts, loss of functional tracheids happens through embolism, which can be partially mitigated by increasing the hydraulic safety of the xylem. However, higher hydraulic safety is usually achieved by building narrower tracheids with thicker walls, resulting in a reduction of the hydraulic efficiency of the xylem (conductivity per unit area). Reduced efficiency constrains water transport, limits photosynthesis and might delay recovery after the drought. Supporting existing research on safety-efficiency tradeoff, we test the hypothesis that under dry conditions, isohydric pinions grow xylem that favor efficiency over safety. Using a seven-year experiment with three watering treatments (drought, control, irrigated) in New Mexico, we investigate the effect of drought on the xylem anatomy of pinions’ branches. We also compare the treatment effect with interannual variations in xylem structure. We measure anatomical variables – conductivities, cell wall thicknesses, hydraulic diameter, cell reinforcement and density – and preliminarily conclude that treatment has little effect on hydraulic efficiency while hydraulic safety is significantly reduced under dry conditions. Taking advantage of an extremely dry year occurrence during the experiment, we find a sharp increase in vulnerability for xylem tissues built the same year.

_____________________________________________________________________

H13S-01: The Water Security Hydra (Invited)

Upmanu Lall

Monday, December 11, 2017 13:40-13:55

New Orleans Ernest N. Morial Convention Center – 280-282

 

As the editor of a new journal on water security, I have been pondering what it can mean theoretically and practically. At one level, it is pretty aobvious that it refers to the ability to affordably and reliably access water of appropriate quality, and to be protected from the water related ravages of nature, such as floods, droughts and water borne disease. The concept of water security can apply to a family, a company, a state or globally. Of course, since we value the environment, water security embraces the needs of the environment. Where, we consider economic development or energy production, water security also emerges as a critical factor. So, in short it touches almost all things about water that pertain to our lives. New stresses are created by a changing climate, growing populations and an ever changing society, economic activity and environment. Thus, if assuring water security is a goal at any of the scales of interest, many factors need to be considered, and what can really be assured, where and for how long emerges as an interesting question. Local (place, time, individuals, politics) as well as global (climate, economics, hydrology) factors interact to determine outcomes, not all of which are readily mapped in our mathematical or cognitive models to a functional notion of what constitutes security in the face of changing conditions and actors. Further, assurance implies going beyond characterization to developing actions, responses to stressors and risk mitigation strategies. How these perform in the short and long run, and what are the outcomes and strategies for impact mitigation in the event of failure then determines water security. Recognizing that providing assurance of water security has always been the goal of water management, regulation and development, perhaps the challenge is to understand what this means from the perspective of not just the “water managers” but the individuals who are the unwitting beneficiaries, or the instruments for the approval of the strategies that are implemented.

In this talk, I will strive to lay out a cognitive framework for how performance evaluation of water security, and instrument design for assurance can be approached from a multi-stress and multi-user perspective. Selected examples will be used to lillustrate the idea in the context of America’s Water.

_____________________________________________________________________

H13M-06: Regionally Strong Feedbacks between the Atmosphere and Terrestrial Biosphere (Invited)

Julia Green, Pierre Gentine

Monday, 11 December 2017  14:55 – 15:10

New Orleans Ernest N. Morial Convention Center – 283-285

 

Vegetation variability modulates water and energy fluxes to the atmosphere with the potential to impact climate and weather patterns that in turn regulate vegetation dynamics. In this study, we quantify variations in the strength of biosphere–atmosphere feedbacks (influencing the hydrologic cycle) across different biomes and timescales and evaluate the ability of Earth System Models to capture them. We use remote sensing data (using Solar Induced Fluorescence as a proxy for photosynthesis) combined with a statistical Multivariate Granger Causality technique to evaluate the feedback strength and the timescale in which they occur, which is then used as a benchmark for model assessment. Our conclusions have the potential to improve climate and weather predictions and provide insight of ecohydrological processes that have regional scale impact (Green, J.K. et al. 2017).

_____________________________________________________________________

H13O-08: A Strategy for a Parametric Flood Insurance Using Proxies

Masahiko Haraguchi, Upmanu Lall

Monday, 11 December 2017 15:25 – 15:40

New Orleans Ernest N. Morial Convention Center – 293-294

 

Traditionally, the design of flood control infrastructure and flood plain zoning require the estimation of return periods, which have been calculated by river hydraulic models with rainfall-runoff models. However, this multi-step modeling process leads to significant uncertainty to assess inundation. In addition, land use change and changing climate alter the potential losses, as well as make the modeling results obsolete. For these reasons, there is a strong need to create parametric indexes for the financial risk transfer for large flood events, to enable rapid response and recovery. Hence, this study examines the possibility of developing a parametric flood index at the national or regional level in Asia, which can be quickly mobilized after catastrophic floods. Specifically, we compare a single trigger based on rainfall index with multiple triggers using rainfall and streamflow indices by conducting case studies in Bangladesh and Thailand. The proposed methodology is 1) selecting suitable indices of rainfall and streamflow (if available), 2) identifying trigger levels for specified return periods for losses using stepwise and logistic regressions, 3) measuring the performance of indices, and 4) deriving return periods of selected windows and trigger levels. Based on the methodology, actual trigger levels were identified for Bangladesh and Thailand. Models based on multiple triggers reduced basis risks, an inherent problem in an index insurance. The proposed parametric flood index can be applied to countries with similar geographic and meteorological characteristics, and serve as a promising method for ex-ante risk financing for developing countries. This work is intended to be a preliminary work supporting future work on pricing risk transfer mechanisms in ex-ante risk finance.

_____________________________________________________________________

GC14A-06: Investigating the Impact of Land-Atmosphere Interactions on the Interannual Carbon Flux

Julia K. Green, Pierre Gentine

Monday, 11 December 2017  17:15 – 17:30

New Orleans Ernest N. Morial Convention Center – 260-262

 

The terrestrial biosphere is the largest contributor to the variability found in the interannual global carbon flux as a result of varying photosynthesis and respiration rates based on climactic conditions and extremes. Due to its large contribution to the carbon cycle, understanding the influence of land-atmosphere feedbacks on the present-day and future CO2 flux is essential to mitigating the risks and hazards associated with rising atmospheric CO2 levels. In this study we use GLACE-CMIP5 model runs1 to understand the impact of land-atmosphere interactions on the global carbon cycle interannually and aim to distinguish the impacts of (1) soil moisture interannual variability, (2) soil moisture long-term trends, and (3) CO2 fertilization in the coming decades. These results can be used to understand what is controlling the net biosphere CO2 flux and to determine whether the terrestrial biosphere will remain a carbon sink in the future.

_____________________________________________________________________

H21M-01: Coupling between the continental carbon and water cycles (Invited)

Pierre Gentine, Leo Adrien Lemordant, Julia K. Green

Tuesday, 12 December 2017  08:00 – 08:15

New Orleans Ernest N. Morial Convention Center – 295-296

 

The continental carbon and water cycles are fundamentally coupled through leaf gas exchange at the stomata level. In this presentation, we will emphasize the importance of this coupling for the future of the water cycle (runoff, evaporation, soil moisture) and in turn the implications for the carbon cycle and the capacity of continents to act as a carbon dioxide sink in the future. Opportunities from coupled carbon-water monitoring platforms will be then emphasized.

_____________________________________________________________________

H22G: Water Data Drought: Addressing Limited Water Demand Data for Research and Policy I

Maura Allaire, Laureline Josset, Upmanu Lall

Tuesday, 12 December 2017  10:20 – 12:20

New Orleans Ernest N. Morial Convention Center – 280-282
Forecasting water demand within key sectors as a function of consumer preferences, climate, and water price is critical for any assessment of sustainability or resilience of water, energy and food systems. Innovation in demand forecasting methods has been limited by the lack of effective data collection and access to such data. Furthermore, the U.S. lags behind other countries concerning an understanding of demand at a national level.

How can we address this water data challenge? Emerging technologies and development of web-based data platforms provide a broad range of possibilities to transform data availability as well as identify strategic policy and investment directions.

The panelists will draw on their extensive experience with water demand across sectors to lay out possible strategies to address this situation. The panel will to discuss key data gaps, innovative approaches for demand monitoring and forecasting, institutional arrangements for data collection and distribution, and emerging technologies.

_____________________________________________________________________

H22B-02: Designing and operating infrastructure for nonstationary flood risk management

James Doss-Gollin, David J Farnham, Upmanu Lall

Tuesday, 12 December 2017  10:35 – 10:50

New Orleans Ernest N. Morial Convention Center – 293-294

 

Climate variability on time scales ranging from several years to several decades causes the risk associated with climate extremes such as floods and droughts to vary in time. We use reconstructed estimates of streamflow spanning several centuries to ground a theoretical study of infrastructure design and operation. Because climate variability on time scales beyond a decade may not be well represented in the historical record, we assess how the length of the historical record (which can be used for estimating climate risk) impacts the reliability of these estimates. We also use mathematical and statistical models to assess how the lifestan of a risk management project impacts probabilistic estimates of this risk. Our methodology can be applied to the problem of sizing new infrastructure in a changing climate, and to the problem of simultaneously comparing long-term and short-term projects for climate risk adaptation.

_____________________________________________________________________

A23L-06: Is evaporative colling important for shallow clouds?

Pierre Gentine, Seung-Bu Park

Tuesday, 12 December 2017 14:45 – 15:00

New Orleans Ernest N. Morial Convention Center – 398-399

 

We investigate and test using large-eddy simulations the hypothesis that evaporative cooling might not be crucial for shallow clouds.

Results from various Shallow convection and stratocumulus LES experiments show that the influence of evaporative cooling is secondary compared to turbulent mixing, which dominates the buoyancy reversal. In shallow cumulus subising shells are not due to evaporative cooling but rather reflect a vortical structure, with a positive buoyancy anomaly in the core due to condensation. Disabling evaporative cooling has negligible impact on this vortical structure and on buoyancy reversal. Similarly, in non-precipitating stratocumuli evaporative cooling is negligible compared to other factors, especially turbulent mixing and pressure effects. These results emphasize that it may not be critical to include evaporative cooling in parameterizations of shallow clouds and that it does not alter entrainment.

_____________________________________________________________________

NH24A-03: Designing Financial Instruments for Rapid Flood Response Using Remote Sensed and Archival Hazard and Exposure Information (Invited)

Upmanu Lall, Maura Allaire, Masahiko Haraguchi

Tuesday, 12 December 2017 16:25 – 16:35

New Orleans Ernest N. Morial Convention Center – 243-244

 

Catastrophic floods can pose a significant challenge for response and recovery. A key bottleneck in the speed of response is the availability of funds to a country or regions finance ministry to mobilize resources. Parametric instruments, where the release of funs is tied to the exceedance of a specified index or threshold, rather than to loss verification are well suited for this purpose. However, designing and appropriate index, that is not subject to manipulation and accurately reflects the need is a challenge, especially in developing countries which have short hydroclimatic and loss records, and where rapid land use change has led to significant changes in exposure and hydrology over time. The use of long records of rainfall from climate re-analyses, flooded area and land use from remote sensing to design and benchmark a parametric index considering the uncertainty and representativeness of potential loss is explored with applications to Bangladesh and Thailand. Prospects for broader applicability and limitations are discussed.

_____________________________________________________________________

PA24A-04: Community Response to Impaired Drinking Water Quality: Evidence from Bottled Water Sales

Maura Allaire, Shuyan Zheng, Upmanu Lall

Tuesday, 12 December 2017  16:45 – 17:00

New Orleans Ernest N. Morial Convention Center – 255-257

 

Drinking water contaminants pose a harm to public health. When confronted with elevated contaminate levels, individuals can take averting actions to reduce exposure, such as bottled water purchases. This study addresses a problem of national interest given that 9 to 45 million people have been affected by drinking water quality violations in each of the past 34 years. Moreover, few studies address averting behavior and avoidance costs due to water quality violations. This study assesses how responses might differ across baseline risk of impaired water quality and demographics of service area.

We match a panel of weekly supermarket sales data with geocoded violations data for 67 counties in the Southeast from 2006-2015. We estimate the change in bottled water sales due to drinking water violations using a fixed effects model. Observing market behavior also allows us to calculate the cost of these averting actions.

Critical findings from this study contribute to understanding how communities respond to water quality violations. We find that violations have considerable effects on bottled water consumption. Sales increase 8.1 percent due to violations related to microorganisms and 31.2 percent due to Tier 1 violations, which pose an immediate health risk.

In addition, we calculate a national cost of averting actions of $26 million for microorganism violations from 2006-2015, which represents a lower-bound estimate. Averting costs vary considerably across the U.S. and some counties bear a particularly large burden, such as in California and Texas.

Overall, this study provides insight into how averting behavior differs across contaminant type, water utility characteristics, and community demographics. Such knowledge can aid public health agencies, water systems, and environmental regulators to direct assistance to communities most in need.

_____________________________________________________________________

PA24A-08: Water data in US: a spatial, temporal and sectoral analysis

Laureline Josset, Maura Allaire, Upmanu Lall

Tuesday, 12 December 2017  17:45 – 18:00

New Orleans Ernest N. Morial Convention Center – 255-257

 

Water data plays a crucial role in the development and implementation of sustainable water management strategies. Both effective design and assessment hinge on accurate information. This requires environmental, climatic, hydrologic, hydrogeologic, industrial, agricultural, energetic and socio-economic data to accurately characterize and project future supply and demand.

In 2001, Vorosmarty et al. painted a stark future for water data, which was qualified as “a new endangered species”. Sixteen years after this publication, we propose a review of the current state of water data in the United States. While considerable progress has been made in data science and model development in the recent years, models are only as good as the data that populate them.

After a brief overview of water data aggregated at the national level, we compare datasets from federal agencies with water information collected by individual states. We note in particular the potential gaps in the collected information that would support research beyond water balance accounts to informing regulations, investments, and economic decisions. In addition, we assess the information structures that host and disseminate data as well as data availability and usability (i.e. whether tools are proposed such as metrics, visualization, projections).

We conclude our paper with a review of the current technological developments, policies and initiatives that may be transformative and redefine the future of water data. We follow two angles: the progress made in data collection (e.g. remote sensing, datascience, reporting policies) and in data dissemination (frameworks, cyber-infrastructures and standards). We review in particular the current initiatives taking place in US and around the world that promote water data freely available to all.

_____________________________________________________________________

NG31B: Urban and Geophysical System Interactions: Did We Learn from Katrina and Other Events? (Part II Posters)

Upmanu Lall

Wednesday, 13 December 2017 08:00 – 12:20

New Orleans Ernest N. Morial Convention Center – Poster Hall D-F

 

This co-organized session will investigate the complex and bidirectional interactions between urban and geophysical systems. Urban systems with their complex mobility, energy and communications networks are a major driver of climate change, and will be crucial both for mitigation and for adaptation. At the same time, most the infrastructure and wealth is concentrated in cities. Katrina and other catastrophic events remind us that many cities are under geophysical threats potentially exacerbated by climate change. Did we really learn from from these events? How far did we shift from traditional resources and risks management to integrated environmental monitoring and responses over a wide range of space-time scales to both ongoing geophysical changes and urbanisation processes? How do we intend to increase the city’s resilience? What are the scientific deadlocks and challenges?

_____________________________________________________________________

NG41B: Urban and Geophysical System Interactions: Did We Learn from Katrina and Other Events? (Part I)

Upmanu Lall

Thursday, 14 December 2017 08:00 – 10:00

New Orleans Ernest N. Morial Convention Center- 228-230

 

This co-organized session will investigate the complex and bidirectional interactions between urban and geophysical systems. Urban systems with their complex mobility, energy and communications networks are a major driver of climate change, and will be crucial both for mitigation and for adaptation. At the same time, most the infrastructure and wealth is concentrated in cities. Katrina and other catastrophic events remind us that many cities are under geophysical threats potentially exacerbated by climate change. Did we really learn from from these events? How far did we shift from traditional resources and risks management to integrated environmental monitoring and responses over a wide range of space-time scales to both ongoing geophysical changes and urbanisation processes? How do we intend to increase the city’s resilience? What are the scientific deadlocks and challenges?

_____________________________________________________________________

GC31D-1032: Optimization of Water Resources and Agricultural Activities for Economic Benefit in Colorado

Joohye Lim, Upmanu Lall

Wednesday, 13 December 2017  08:00 – 12:20 New Orleans Ernest N. Morial Convention Center – Poster Hall D-F

 

The limited water resources available for irrigation are a key constraint for the important agricultural sector of Colorado’s economy. As climate change and groundwater depletion reshape these resources, it is essential to understand the economic potential of water resources under different agricultural production practices. This study uses a linear programming optimization at the county spatial scale and annual temporal scales to study the optimal allocation of water withdrawal and crop choices. The model, AWASH, reflects streamflow constraints between different extraction points, six field crops, and a distinct irrigation decision for maize and wheat. The optimized decision variables, under different environmental, social, economic, and physical constraints, provide long-term solutions for ground and surface water distribution and for land use decisions so that the state can generate the maximum net revenue. Colorado, one of the largest agricultural producers, is tested as a case study and the sensitivity on water price and on climate variability is explored.

_____________________________________________________________________

NH33B-0257: Assessing Risks of Mine Tailing Dam Failures

Paulina Concha Larrauri, Upmanu Lall

Wednesday, 13 December 2017  13:40 – 18:00

New Orleans Ernest N. Morial Convention Center – Poster Hall D-F

 

The consequences of tailings dam failures can be catastrophic for communities and ecosystems in the vicinity of the dams. The failure of the Fundão tailings dam at the Samarco mine in 2015 killed 19 people with severe consequences for the environment. The financial and legal consequences of a tailings dam failure can also be significant for the mining companies. For the Fundão tailings dam, the company had to pay 6 billion dollars in fines and twenty-one executives were charged with qualified murder. There are tenths of thousands of active, inactive, and abandoned tailings dams in the world and there is a need to better understand the hazards posed by these structures to downstream populations and ecosystems.

A challenge to assess the risks of tailings dams in a large scale is that many of them are not registered in publicly available databases and there is little information about their current physical state. Additionally, hazard classifications of tailings dams – common in many countries- tend to be subjective, include vague parameter definitions, and are not always updated over time. Here we present a simple methodology to assess and rank the exposure to tailings dams using ArcGIS that removes subjective interpretations. The method uses basic information such as current dam height, storage volume, topography, population, land use, and hydrological data. A hazard rating risk was developed to compare the potential extent of the damage across dams. This assessment provides a general overview of what in the vicinity of the tailings dams could be affected in case of a failure and a way to rank tailings dams that is directly linked to the exposure at any given time. One hundred tailings dams in Minas Gerais, Brazil were used for the test case. This ranking approach could inform the risk management strategy of the tailings dams within a company, and when disclosed, it could enable shareholders and the communities to make decisions on the risks they are taking.

_____________________________________________________________________

NH33B-0246: Assessing the adequacy of water storage infrastructure capacity under hydroclimatic variability and water demands in the United States

Michelle Ho, Upmanu Lall

Wednesday, 13 December 2017  13:40 – 18:00

New Orleans Ernest N. Morial Convention Center – Poster Hall D-F

 

As populations and associated economic activity in the US evolve, regional demands for water likewise change. For regions dependent on surface water, dams and reservoirs are critical to storing and managing releases of water and regulating the temporal and spatial availability of water in order to meet these demands. Storage capacities typically range from seasonal storage in the east to multi-annual and decadal-scale storage in the drier west. However, most dams in the US were designed with limited knowledge regarding the range, frequency, and persistence of hydroclimatic extremes. Demands for water supplied by these dams have likewise changed. Furthermore, many dams in the US are now reaching or have already exceeded their economic design life.

The converging issues of aging dams, improved knowledge of hydroclimatic variability, and evolving demands for dam services result in a pressing need to evaluate existing reservoir capacities with respect to contemporary water demands, long term hydroclimatic variability, and service reliability into the future. Such an effort is possible given the recent development of two datasets that respectively address hydroclimatic variability in the conterminous United States over the past 555 years and human water demand related water stress over the same region. The first data set is a paleoclimate reconstruction of streamflow variability across the CONUS region based on a tree-ring informed reconstruction of the Palmer Drought Severity Index. This streamflow reconstruction suggested that wet spells with shorter drier spells were a key feature of 20th century streamflow compared with the preceding 450 years. The second data set in an annual cumulative drought index that is a measure of water balance based on water supplied through precipitation and water demands based on evaporative demands, agricultural, urban, and industrial demands. This index identified urban and regional hotspots that were particularly dependent on water transfers and vulnerable to persistent drought risk. These data sets are used in conjunction with the national inventory of dams to assess the current capacity of dams to meet water demands considering variability in streamflow over the past 555 years. A case study in the North-East US is presented.

_____________________________________________________________________

NG33A-0180: Application of Deep Learning and Supervised Learning Methods to Recognize Nonlinear Hidden Pattern in Water Stress Levels from Spatiotemporal Datasets across Rural and Urban US Counties

Tristan Eisenhart, Laureline Josset, Upmanu Lall

Wednesday, 13 December 2017  13:40 – 18:00

New Orleans Ernest N. Morial Convention Center – Poster Hall D-F

 

In the wake of recent water crises, the need to understand and predict the risk of water stress in urban and rural areas has grown. This understanding has the potential to improve decision making in public resource management, policy making, risk management and investment decisions.

Assuming an underlying relationship between urban and rural water stress and observable features, we apply Deep Learning and Supervised Learning models to uncover hidden nonlinear patterns from spatiotemporal datasets. Results of interest includes prediction accuracy on extreme categories (i.e. urban areas highly prone to water stress) and not solely the average risk for urban or rural area, which adds complexity to the tuning of model parameters.

We first label urban water stressed counties using annual water quality violations and compile a comprehensive spatiotemporal dataset that captures the yearly evolution of climatic, demographic and economic factors of more than 3,000 US counties over the 1980-2010 period. As county-level data reporting is not done on a yearly basis, we test multiple imputation methods to get around the issue of missing data. Using Python libraries, TensorFlow and scikit-learn, we apply and compare the ability of, amongst other methods, Recurrent Neural Networks (testing both LSTM and GRU cells), Convolutional Neural Networks and Support Vector Machines to predict urban water stress. We evaluate the performance of those models over multiple time spans and combine methods to diminish the risk of overfitting and increase prediction power on test sets.

This methodology seeks to identify hidden nonlinear patterns to assess the predominant data features that influence urban and rural water stress. Results from this application at the national scale will assess the performance of deep learning models to predict water stress risk areas across all US counties and will highlight a predominant Machine Learning method for modeling water stress risk using spatiotemporal data.

_____________________________________________________________________

NH41E: Wicked Problems and Deep Uncertainties: Tough Challenges in Natural Hazard Assessment and Mitigation I

Michael S Steckler, Upmanu Lall

Thursday, 14 December 2017 08:00 – 10:00

New Orleans Ernest N. Morial Convention Center – 245

 

Natural disasters claim lives and cause billions of dollars of losses around the world. Assessing and mitigating these hazards challenges societies with “wicked” problems, in which crucial information is missing and proposed solutions involve complex interactions with other societal goals. For example: How should we prepare for a great earthquake in areas where tectonics favor such events but we have no evidence that they have occurred and hence how large they may be or how often to expect them? How can we assess the hazard when the recurrence of large earthquakes, floods, or hurricanes is changing with time or is expected to do so? This session delves into the scientific and business issues, utilizing state‐of‐the‐art methods and advanced analytics for better understanding, quantification, and management of natural hazards and risks. We invite contributions from disciplines including hydrology, atmospheric sciences, seismology, computer sciences, machine learning, insurance and risk management.

_____________________________________________________________________

NH41B: Wicked Problems and Deep Uncertainties: Tough Challenges in Natural Hazard Assessment and Mitigation II Posters

Michael S Steckler, Upmanu Lall

Thursday, 14 December 2017 08:00 – 12:20

New Orleans Ernest N. Morial Convention Center – Poster Hall D-F

 

Natural disasters claim lives and cause billions of dollars of losses around the world. Assessing and mitigating these hazards challenges societies with “wicked” problems, in which crucial information is missing and proposed solutions involve complex interactions with other societal goals. For example: How should we prepare for a great earthquake in areas where tectonics favor such events but we have no evidence that they have occurred and hence how large they may be or how often to expect them? How can we assess the hazard when the recurrence of large earthquakes, floods, or hurricanes is changing with time or is expected to do so? This session delves into the scientific and business issues, utilizing state‐of‐the‐art methods and advanced analytics for better understanding, quantification, and management of natural hazards and risks. We invite contributions from disciplines including hydrology, atmospheric sciences, seismology, computer sciences, machine learning, insurance and risk management.

_____________________________________________________________________

A41M: Land-Atmosphere Interactions, Biosphere-Boundary Layer Feedbacks, and Moist Convection III

Pierre Gentine

Thursday, 14 December 2017  08:00 – 10:00

New Orleans Ernest N. Morial Convention Center – 398-399

 

Interactions and feedbacks at the soil-vegetation-atmospheric boundary layer interface crucially affect biological, chemical, and physical ecosystem processes such as surface-energy partitioning, evapotranspiration, and the exchange of atmospheric trace gases. These processes in turn impact boundary-layer structures and growth, clouds, and convective development. An improved understanding of these feedbacks can help reduce uncertainties associated with global energy and water cycles, land-management, and the climate system. This session is open to theoretical, experimental, remote sensing, and modeling studies in land-atmosphere interactions across natural and managed ecosystems, from the local to regional scale. We appreciate contributions explore the role of both vegetation and atmosphere in modulating energy and matte exchange within the coupled surface-atmosphere system. Of special interest are novel approaches that fuse observations and models or that capture land-atmosphere feedbacks the across multiple spatial and temporal scales.

_____________________________________________________________________

A42D: Land-Atmosphere Interactions, Biosphere-Boundary Layer Feedbacks, and Moist Convection II

Pierre Gentine

Thursday, 14 December 2017 10:20 – 12:20

New Orleans Ernest N. Morial Convention Center – 398-399

 

Interactions and feedbacks at the soil-vegetation-atmospheric boundary layer interface crucially affect biological, chemical, and physical ecosystem processes such as surface-energy partitioning, evapotranspiration, and the exchange of atmospheric trace gases. These processes in turn impact boundary-layer structures and growth, clouds, and convective development. An improved understanding of these feedbacks can help reduce uncertainties associated with global energy and water cycles, land-management, and the climate system. This session is open to theoretical, experimental, remote sensing, and modeling studies in land-atmosphere interactions across natural and managed ecosystems, from the local to regional scale. We appreciate contributions explore the role of both vegetation and atmosphere in modulating energy and matter exchange within the coupled surface-atmosphere system. Of special interest are novel approaches that fuse observations and models or that capture land-atmosphere feedbacks the across multiple spatial and temporal scales.

_____________________________________________________________________

PP42A-03: Inferring Spatio-temporal Variations in the Risk of Extreme Precipitation in the Western United States from Tree-ring Chronologies

Scott Steinschneider, Michelle Ho, Upmanu Lall

Thursday, 14 December 2017  10:50 – 11:05

New Orleans Ernest N. Morial Convention Center – 344-345

 

This work explores how extreme cold-season precipitation dynamics along the west coast of the United States have varied in the past under natural climate variability through an analysis of the moisture anomalies recorded by tree-ring chronologies across the coast and interior of the western U.S. Winters with high total precipitation amounts in the coastal regions are marked by a small number of extreme storms that exhibit distinct spatial patterns of precipitation across the coast and further inland. Building from this observation, this work develops a novel application of dendroclimatic evidence to explore the following questions: a) how is extreme precipitation variability expressed in a network of tree-ring chronologies; b) can this information provide insight on the space-time variability of storm tracks that cause these extreme events; and c) how can the joint variability of extreme precipitation and storm tracks be modeled to develop consistent, multi-centennial reconstructions of both? We use gridded, tree-ring based reconstructions of the summer Palmer Drought Severity Index (PDSI) extending back 500 years within the western U.S. to build and test a novel statistical framework for reconstructing the space-time variability of coastal extreme precipitation and the associated wintertime storm tracks. Within this framework, we (1) identify joint modes of variability of extreme precipitation fields and tree-ring based PDSI reconstructions; (2) relate these modes to previously identified, unique storm track patterns associated with atmospheric rivers (ARs), which are the dominant type of storm that is responsible for extreme precipitation in the region; and (3) determine latitudinal variations of landfalling ARs across the west coast and their relationship to the these joint modes. To our knowledge, this work is the first attempt to leverage information on storm track patterns stored in a network of paleoclimate proxies to improve reconstruction fidelity.

_____________________________________________________________________

H42B-06: Vegetation physiology controls continental water cycle responses to climate change

Leo Adrien Lemordant, Jacob Scheff, Pierre Gentine

Thursday, 14 December 2017  11:35 – 11:50

New Orleans Ernest N. Morial Convention Center – 293-294

 

Predicting how climate change will affect the hydrologic cycle is of utmost importance for ecological systems and for human life and activities. A typical perspective is that global warming will cause an intensification of the mean state, the so-called “dry gets drier, wet gets wetter” paradigm. While this result is robust over the oceans, recent works suggest it may be less appropriate for terrestrial regions. Using Earth System Models (ESMs) with decoupled surface (vegetation physiology, PHYS) and atmospheric (radiative, ATMO) CO2responses, we show that the CO2 physiological response dominates the change in the continental hydrologic cycle compared to radiative and precipitation changes due to increased atmospheric CO2, counter to previous assumptions. Using multiple linear regression analysis, we estimate the individual contribution of each of the three main drivers, precipitation, radiation and physiological CO2 forcing (see attached figure). Our analysis reveals that physiological effects dominate changes for 3 key indicators of dryness and/or vegetation stress (namely LAI, P-ET and EF) over the largest fraction of the globe, except for soil moisture which exhibits a more complex response. This highlights the key role of vegetation in controlling future terrestrial hydrologic response.

_____________________________________________________________________

H44C-06: When Does Vapor Pressure Deficit Drive or Reduce Evaporation?

Adam Massmann, Pierre Gentine

Thursday, 14 December 2017  17:15 – 17:30

New Orleans Ernest N. Morial Convention Center – 293-294

 

Depending on plant response (e.g. stomatal closure), ecosystem-scale evaporation can either increase or decrease with changes in vapor pressure deficit. This ecosystem response drives evaporation and atmospheric moisture feedbacks. We use data from 75 FluxNet sites within a Penman-Monteith framework to examine when ecosystem evaporation is suppressed or enhanced by increases in vapor pressure deficit. Evaporation response is quantified as a function of soil moisture, atmospheric conditions, and plant functional type. Uncertainty in plant response is accounted for by varying the stomatal resistance model and its parameters. This in-situ observation-based analysis aids understanding for how ecosystems will respond and/or contribute to future shifts in atmospheric water demand.

_____________________________________________________________________

GC44A-07: Predicting summer residential electricity demand across the U.S.A using climate information

Xun Sun, Siyan Wang, Upmanu Lall

Thursday, 14 December 2017  17:30 – 17:45

New Orleans Ernest N. Morial Convention Center – 260-262

 

We developed a Bayesian Hierarchical model to predict monthly residential per capita electricity consumption at the state level across the USA using climate information. The summer period was selected since cooling requirements may be directly associated with electricity use, while for winter a mix of energy sources may be used to meet heating needs. Historical monthly electricity consumption data from 1990 to 2013 were used to build a predictive model with a set of corresponding climate and non-climate covariates. A clustering analysis was performed first to identify groups of states that had similar temporal patterns for the cooling degree days of each state. Then, a partial pooling model was applied to each cluster to assess the sensitivity of monthly per capita residential electricity demand to each predictor (including cooling-degree-days, gross domestic product (GDP) per capita, per capita electricity demand of previous month and previous year, and the residential electricity price). The sensitivity of residential electricity to cooling-degree-days has an identifiable geographic distribution with higher values in northeastern United States.

_____________________________________________________________________

A51C: Land-Atmosphere Interactions, Biosphere-Boundary Layer Feedbacks, and Moist Convection I Posters

Pierre Gentine

Friday, 15 December 2017 08:00 – 12:20

New Orleans Ernest N. Morial Convention Center – Poster Hall D-F

 

Interactions and feedbacks at the soil-vegetation-atmospheric boundary layer interface crucially affect biological, chemical, and physical ecosystem processes such as surface-energy partitioning, evapotranspiration, and the exchange of atmospheric trace gases. These processes in turn impact boundary-layer structures and growth, clouds, and convective development. An improved understanding of these feedbacks can help reduce uncertainties associated with global energy and water cycles, land-management, and the climate system. This session is open to theoretical, experimental, remote sensing, and modeling studies in land-atmosphere interactions across natural and managed ecosystems, from the local to regional scale. We appreciate contributions explore the role of both vegetation and atmosphere in modulating energy and matte exchange within the coupled surface-atmosphere system. Of special interest are novel approaches that fuse observations and models or that capture land-atmosphere feedbacks the across multiple spatial and temporal scales.

_____________________________________________________________________

 

H51I-1387: Zonal wind indices to reconstruct United States winter precipitation during El Nino

David Farnham, Scott Steinschneider, Upmanu Lall

Friday, 15 December 2017  08:00 – 12:20

New Orleans Ernest N. Morial Convention Center – Poster Hall D-F

 

The highly discussed 2015/16 El Nino event, which many likened to the similarly strong 1997/98 El Nino event, led to precipitation impacts over the continental United States (CONUS) inconsistent with general expectations given past events and model-based forecasts. This presents a challenge for regional water managers and others who use seasonal precipitation forecasts who previously viewed El Nino events as times of enhanced confidence in seasonal water availability and flood risk forecasts. It is therefore useful to understand the extent to which wintertime CONUS precipitation during El Nino events can be explained by seasonal sea surface temperature heating patterns and the extent to which the precipitation is a product of natural variability. In this work, we define two seasonal indices based on the zonal wind field spanning from the eastern Pacific to the western Atlantic over CONUS that can explain El Nino precipitation variation spatially throughout CONUS over 11 historic El Nino events from 1950 to 2016. The indices reconstruct El Nino event wintertime (Jan-Mar) gridded precipitation over CONUS through cross-validated regression much better than the traditional ENSO sea surface temperature indices or other known modes of variability. Lastly, we show strong relationships between sea surface temperature patterns and the phases of the zonal wind indices, which in turn suggests that some of the disparate CONUS precipitation during El Nino events can be explained by different heating patterns. The primary contribution of this work is the identification of intermediate variables (in the form of zonal wind indices) that can facilitate further studies into the distinct hydroclimatic response to specific El Nino events.

_____________________________________________________________________

A51A-2014: Test of Monin-Obukhov similarity theory using distributed temperature sensing

Yu Cheng, Qi Li, Pierre Gentine

Friday, 15 December 2017  08:00 – 12:20

New Orleans Ernest N. Morial Convention Center – Poster Hall D-F

 

Monin-Obukhov similarity theory [Monin and Obukhov, 1954] (MOST) has been widely used to calculate atmospheric surface fluxes applying the structure correction functions [Stull, 1988]. The exact forms of the structure correction functions for momentum and heat, which depend on the vertical gradient velocity and temperature, have been determined empirically mostly from the Kansas experiment [Kaimal et al., 1972]. However, due to the limitation of point measurement, the vertical gradient of temperature and horizontal wind speed are not well captured. Here we propose a way to measure the vertical gradient of temperature and horizontal wind speed with high resolution in space (every 12.7 cm) and time (every second) using the Distributed Temperature Sensing [Selker et al., 2006] (DTS), thus determining the exact form of the structure correction functions of MOST under various stability conditions. Two parallel vertical fiber optics will be placed on a tower at the central facility of ARM SGP site. Vertical air temperature will be measured every 12.7 cm by the fiber optics and horizontal wind speed along fiber will be measured. Then vertical gradient of temperature and horizontal wind speed will be calculated and stability correction functions for momentum and heat will be determined.

_____________________________________________________________________

A51A-2036: The implications of turbulent coherent structures on Monin-Obukhov similarity theory in a dry convective boundary layer

Qi Li, Pierre Gentine

Friday, 15 December 2017  08:00 – 12:20

New Orleans Ernest N. Morial Convention Center – Poster Hall D-F

 

According to Townsend’s hypothesis, so-called ‘wall-attached’ eddies are believed to be the main contributors to turbulent transport in the atmospheric surface layer. This is also the cornerstone for one of the assumptions of Monin-Oubkhov similarity theory (MOST). However, previous evidence shows that the outer-scale eddies can interact and impact the surface layer, resulting in deviations from the classic MOST scaling. We conduct large-eddy simulations and direct numerical simulations of a dry convective boundary layer to investigate the impact of buoyancy on coherent structures in the surface layer. The turbulent coherent structures are identified by releasing a height-dependent passive tracer and categorized as updrafts and subsidence. The MOST similarity functions computed from the simulation results indicate a larger deviation of \phi_m from \phi_m~(-z/L)^{1/4} than the corresponding temperature similarity function (\phi_h), consistent with other previous simulations results. Analyses of turbulent coherent structures show that as instability increases, there is a general change in the structure of updrafts and subsidence. Updrafts act as active eddies and are the dominant contributor in the surface layer for different stabilities. Subsidence, which comprises eddies that originate from aloft, contribute increasingly to the transport of temperature but less so for momentum with increasing instability. Such differences can be attributed to the pressure effect on momentum, which results in significant structural difference in the spatial variability of u′ in subsidence. These results demonstrate the mechanism for deviation from MOST might be the involvement of stronger subsidence, relating the possible cause of distinction between and to the difference in turbulent structures.

_____________________________________________________________________

Speaker Sessions:

NH022:

(Un)natural Hazards: Aging US Dams and Levees -Monitoring, Maintenance, Renewal, and Risk

Michelle Ho, Upmanu Lall

 

The US contains more than 30,000 miles of levees and 88,000 dams – 13,500 of these dams are rated as high hazard dams. Safety inspections, monitoring, and restoration varies dramatically by state and ownership. It is clear from the recent failures of spillways, levees, small dams, and tailing mine dams across the country that the state of these aging infrastructure elements is of serious concern.

What are the potential ways to quantify the risks associated with the portfolio of water infrastructure in the US, communicate these risks, and fund management strategies for the future?

The panelists will use their extensive experience with dams and levees over the last century to discuss the issues of quantifying risks associated with the portfolio of water infrastructure in the US, the threats posed by increasing incidence of extreme climate events, how to communicate these risks, and funding strategies for future infrastructure investments.

_____________________________________________________________________

H098:

Pursuing Water Security in Socio-Hydrological Systems

Upmanu Lall

 

Water insecurity is multi-faceted, arising from the challenge of balancing human and environmental water needs together with complex social and political frameworks. Pursuing water security requires understanding of the dynamic feedbacks between humans and water systems at scales from local to planetary, over days to millennia. This session invites studies presenting approaches, concepts, models and datasets that help understand the nature and causes of water (in)security. Submissions are encouraged on all aspects of water security: water scarcity (competition, conflict, collaboration), water pollution and health, flooding (vulnerability, dynamic flood risk), water access (quality, quantity, reliability, equity) as well drivers (infrastructure, economic policy, trade, governance), in addition to water, energy, and food resource availability and distribution. Research on adaptation strategies or solutions to water insecurity, including lessons learned from past or contemporary societies are also invited. Papers will be considered for a special issue of Water Security, a new Elsevier journal.

_____________________________________________________________________

PA036:
Water and Society: The Future of America’s Water: Understanding the landscape of water risks, and addressing the associated societal and economic impacts

James Rising and Mengqian Lu

 

Water is critical for economic growth and security, food and energy production, industrial and urban activities, and environmental services. The United States is currently burdened with aging infrastructure which undermines water quality and exposes communities to the risk of catastrophic failures.  In addition, water policies across the Unites States are often regionally specific and do not consider the hydrologic system, leading to inefficient use and overuse of ground and surface water resources. We seek submissions that address America’s major water challenges and opportunities including the nexus of food-energy-water, the changing landscape of water risks (quality, infrastructure, spatial and temporal availability), the impacts of climate change and variability on policymaking, and methods for addressing these issues into the future on a national level.

_____________________________________________________________________

H082:
Machine Learning Applications in Earth Science and Remote Sensing

Hamed Alemohammad and Pierre Gentine

 

Remote sensing technology has facilitated data acquisition over large and small spatio-temporal scales resulting in valuable large datasets. Present advances in technology with availability of large computing and scalable storage allow us to explore these datasets in a unique way to use data-driven techniques to learn patterns, understand physical processes, and characterize feedbacks in the Earth system. Among the many fields within data exploration, Machine Learning (ML) has had a major impact on not only scientific thinking but also in commercial applications. ML’s advantages include, application flexibility and scaling, fast running time, and ability to represent complex relationships from historical data in conjunction with physics-guided processes. This session is a platform to discuss advancements in the applications of ML in Earth science and remote sensing as well as challenges and future opportunities for ML applications. Presentations will address the merger of discrete mathematics, statistics, physics, and non-linear optimization techniques.

_____________________________________________________________________

NH005:
Dams and Reservoirs – Natural Hazards, Risks, and Solutions

Michelle Ho, Xun Sun, and Upmanu Lall

 

Throughout civilization dams and reservoirs have provided a means of addressing spatial and temporal variability in water supply. Demands for reliable supplies of water for irrigation, hydropower, municipalities, navigation, and flood control result in social and environmental tradeoffs and concerns regarding the safety of dam structures and release patterns. Dam risks vary globally with the US facing aging dam issues, while countries such as China, South Korea, and Japan are constructing new dams and addressing changes in environmental regimes. Identifying and addressing these challenges are the focus of this session. We welcome contributions addressing dam hazards and risks associated with uncontrolled or emergency water releases, changes in flood risk, seismic activity, and increasing age and the need for restoration, retrofitting, or removal. Innovative methods of identifying and monitoring dams and dam safety, assessing safety and cost trade-offs, environmental impacts, sedimentation, and achieving robust dam design and operation are also encouraged.

_____________________________________________________________________

Submit Comment

Your email address will not be published. Required fields are marked *