Dynamic Data-Driven Application Systems


2006 Modeling Sensing, and Environmental Systems Workshop (MoSES 2006)

LNCC, July 10-14, 2006

Presentation Titles and Abstracts

Welcome to the Workshop
Craig Douglas and Maurício Kritz

A brief set of remarks introducing the objectives of the workshop and the format.

Dynamic Data Driven Applications in Porous Media Flow
Richard Ewing
Texas A&M University

The optimization of the production of hydrocarbons from underground reservoirs requires the simulation of processes described by large systems of coupled nonlinear partial differential equations. Due to the uncertainty of the knowledge of the flow properties of the reservoir and the resident fluids, the production measurements from outflow wells must be monitored closely to compare with the data from the simulated processes. The data must be used in real time to make decisions on what fluids to inject, the flow rates of injection and production, the location of new injection or production wells, the utilization of surface facilities, etc. Breakthrough times and concentrations at production wells must be utilized effectively to reduce the bypassing of petroleum in the reservoir and to optimize the production. We discuss a suite of software capabilities and the structure and organization of these codes to control the workflow, change the modules utilized, and match the measured data as effectively as possible. Computational examples from complex reservoir simulations will be presented.

Building Bridges toward Biologists and Ecologists Using DDDAS
Maurício Kritz
LNCC - National Laboratory for Scientific Computing

I will give an informal talk on the following key ideas:

Flux and Transport of Nutrients in the Cantão State Park
Claudia Mazza Dias and Dayse Haime Pastore
LNCC - National Laboratory for Scientific Computing

The use of modeling in the study, project and management in hydrology is not questionable, in the face of the complexity of lakes, reservoirs, rivers and hydrographic basins.

The Cantão State Park is located in Tocantins State (Amazonia) in a delta formed by Javaés, Coco and Araguaia rivers (72 km of length and 12 km of width). It is a wetland where the alternate of weather runs the biogeochemistry cycle. Therefore the idea of this work is to obtain two different models, one for the wet season and other for the dry season. With these models we will try to develop an integrated model for the water flux dynamic for the area of study. Afterwards, it will be used a numerical method to deal with a large number of operations through the use of computers and the amount of memory and processing involved in the process.

Solution Updating in a Dynamic Data Driven Simulation
Yalchin Efendiev, Richard Ewing, Raytcho Lazarov (Texas A&M)
Craig C. Douglas (University of Kentucky and Yale University)
Martin Cole, Greg Jones, Chris R. Johnson (University of Utah)

In this talk we discuss some numerical procedures involved in dynamic data driven simulations (DDDAS) for the contaminant transport in heterogeneous porous media. We consider the contaminant transport in porous media with a number of sensors placed at some locations. One of the objectives of dynamic data driven simulations is to incorporate the sensor data into real time simulations that run continuously. Unlike traditional approaches, in which a static input data set is used as initial conditions only, our approach intends to assimilates many sets of data and corrects computed errors above a given level (which can change during the course of the simulation) as part of the computational process. We have developed and used a number of techniques for dynamic data driven simulations. These techniques are multi-scale interpolation, initial data recovery and permeability sampling. In this talk, we will describe the solution update based on short term past information. Some of our approaches are probabilistic and allow us to assess the uncertainties associated with the predictions.

Sensors, DDDAS, Little Green Men on Mars, and Drugs
Robert Lodder, Clay Harris, and Craig C. Douglas (University of Kentucky)
Gergana Bencheva and Ulrich Langer (Johannes Kepler University of Linz)

Intelligent sensors are key to making DDDAS work. Without them there can be no symbiotic relationship between an application simulation and dynamic steering. A good DDDAS sensor is not only robust, but extremely flexible and reprogrammable on the fly. For example, one can be used either underwater or on Mars. A description of the Solid- State Spectral Imager (SSSI) sensor and how it can be used almost anywhere by almost anyone will be described for applications such are contaminant identification in water bodies, geological identification on a Mars rover will be given. We will briefly give a description, also, of how to deliver drugs to a given spot in a body using DDDAS techniques.

Finslerian systems in forest succession
Peter Antonelli (University of Alberta and UERJ)
Solange Rutz (UERJ)

Use is made of Clement?s classification scheme to model forest succession.An approximation procedure is applied and reduces the forest to 2 species: dominant and codominant.Volterra-Hamilton systems are used to model the dynamics of modular populations, after John Harper.There are 2 time scales,one ecological(real)one and a longer time scale for primary production towards climax.The climax arises as a result of transformations along a sere. These carry one ecoscene into another,perhaps far removed in real time from the first.The main result is that there are 8 types of climax our model allows for dominant/codominant forestry.Each preserves its primary production cost functional. All are Jacobi stable and steady states are linearly stable for various parameter ranges.

Flooded Ecosystems and the Flow of Water
Hélder Lima de Queiroz

The annual rise of waters in the Amazon wetlands induces important changes in the availability of resources for both terrestrial and aquatic animals. We are interested in understanding how these changes affect their life habits and fish communities species composition in the Mamirauá reserve, situated at the confluence of rivers Solimões e Japurá. Preliminary data indicate that the water quality profile changes substantially along the year and within the reserve. We need to know how the water flows in this region to understand how the pulse affect animals, since variations in water flow intensity and water column height throughout the area may affect water composition and whatever is correlated to it.

Dynamic Data-Driven Ocean Models
Mohamed Iskandarani
University of Miami

In this talk an ocean model will be described and how it can be adapted to DDDAS. We will describe a real situation in the Hudson river estuary in which using the SSSI (see Rob Lodder's talk) will enhance our simulations and lead to new discoveries.

Water Quality and DDDAS
Anthony Vodacek
Rochester Institute of Technology

How do you know that the water you are drinking or bathing in will not kill you? Only sensor based testing will provide that answer along with a chemical analysis. What chemicals should be tested for? Only model driven analysis will answer this question. In this talk, the symbiotic nature of water quality and sensors will be discussed.

Estimation of Missing Data
Fabiano Gomes de Oliveira

Missing data problems are an important matter in statistical literature. We discuss some methods to estimate missing values in an incomplete multivariate data set. An application will be shown using rain data form the Amazon region.

Ecology of White-Water Amazonian Floodplain Forests (Vazeas)
Maria Teresa Fernandez Piedade (INPA, Projeto INPA/Max-Planck)
Florian Wittmann (Max-Planck Institute, Projeto INPA/Max-Planck)
Jochen Schoengart (Max-Planck Institute, Projeto INPA/Max-Planck)

Floodplains of the large Amazon rivers cover an area of 6% (300,000 km2) of the Amazon Region, however, adding to this number the mangrove and small streams at terra firma, floodplains areas in the Amazon may cover about 20% of the region. Two different floodplains types can be recognized: the rich varzeá (200,000 km2) and the poor igap? (100,000 km2). Soils and water are compatible different between the two systems. Owing to the higher nutrient status of the varzeá, 90% of the rural population in the Amazon State lives in these areas.

Floodplains are highly dynamic systems with intense processes of inundation and sedimentation. These conditions create different biotic conditions according to the days of inundation per year, leading to high habitats diversity and different plant communities according to their adaptation to oxygen depletion and alcohol intoxication. Succession in the varzeá starts with well adapted and productive aquatic macrophytes, with productivities varying from 6 to 100 t/ha/year, according to the species and time available for production. Trees show different strategies to cope with the flooding leading to zonation of species along the gradient with defined forest types, differing in species composition, diversity, stand density and forest architecture. Different varzeá forest types can be classified as different successional stages and can be recognized by remote sensing techniques which allow detection and area calculation by supervised classification (LandsatTM).

Growth in Amazon floodplains forests is triggered by the monomodal flood-pulse leading to reduction of cambial activity, and formation of annual rings. Despite this reduction in growth, and the relatively low aboveground wood biomass accumulation, varzeá forests are characterized by higher aboveground wood biomass production than the terra firma forests. Tree species diversity of white-water floodplain forests increases along the following gradients: With increasing stand age (succession); With decreasing mean flood-level; With increasing latitude (Wittmann et al., 2006). More than 1000 tree species occur in the varzeá floodplain forests, 60% from those are endemic.

Demonstrating the Validity of a Wildfire DDDAS
Janice Coen (National Center for Atmospheric Research)
Jan Mandel and Jonathan Beezley (University of Colorado at Denver)
Anthony Vodacek and Robert Kremens (Rochester Institute of Technology)
Guan Qin (Texas A&M)
Craig C. Douglas (University of Kentucky)

We report on an ongoing effort to build a Dynamic Data Driven Application System (DDDAS) for short-range forecast of weather and wildfire behavior from real-time weather data, images, and sensor streams. The system changes the forecast as new data is received. We encapsulate the model code and apply an ensemble Kalman filter in time-space with a~highly parallel implementation. In this paper, we discuss how we will demonstrate that our system works using a DDDAS testbed approach and data collected from an earlier fire.

Dynamic Data-Driven Application and SCIRun
Chris R. Johnson
University of Utah

One of the significant challenges for DDDAS is to create software infrastructure and tools that help DDDAS researchers tackle the multidisciplinary, often large-scale, dynamically coupled problems described in the previous presentation.

DDDAS problems often require using multiple software frameworks and packages, which leads to the significant software architecture challenge of integrating and providing interoperability of different software frameworks, packages, and libraries. Our approach to this challenge is to create software "bridges" using a meta-component model that allows the user to easily connect one software framework or package to another.

The new system (currently called SCIRun2, but that will change very soon) support the entire life cycle of scientific applications by allowing scientific programmers to quickly and easily develop new techniques, debug new implementations, and apply known algorithms to solve novel problems. SCIRun2 also contain many powerful visualization algorithms for scalar, vector, and tensor field visualization, as well as image processing tools.

In this presentation, we will provide examples of DDDAS software integration.

Transport of Nutrients and Contaminants in Amazon Rivers
Renato S. Silva
LNCC - National Laboratory for Scientific Computing

We present two set of models statistical and numerical, that are in an initial stage, about nutrient and contaminant transport in rivers. They reflect human impact in the environment and are to be used as a tool to help the administrative decisions. Some results are presented with respect to Amazon Basin rivers, particularly for prevision for discharge, temperature and dissolved oxygen.

When DDDAS Meets HPC: A Train Wreck in the Making?
Craig C. Douglas
University of Kentucky

DDDAS enabled applications run in a different manner than many traditional applications. They place new and quite different strains on high performance systems and centers. A few computer vendors, such as IBM, have woken up and are addressing some of the strains. Others are hoping the emerging field will just go away, which is not likely. This talk will categorize a number of DDDAS-HPC issues and begin to address solutions.

A Seasonal Mathematical Model For Malaria Spreading With Partial Health Care
Luiz Bevilacqua and Ana Paula Wise

Malaria spreading in the Amazonian region in order to be more representative must introduce seasonal effects. Mosquitoes population density varies along the year depending on the precipitation intensity. During the raining season eggs and larvae are washed away and the population density decreases. During the dry season breeding conditions are favorable and the population density increases. This behavior induces a corresponding fluctuation on the number of infection cases reported. If however health care is introduced malaria spreading can come under control. The interesting result is that not all population must receive a rigorous treatment in order to reach a progressive reduction in the number of cases. In the best case only 20% to 40% of the population needs to receive very good health care to reduce the number of cases reported. This pattern, however, depends on temperature. For high temperatures the number of people under rigorous treatment must increase to obtain satisfactory results. Again global warming comes in with negative effects.

This website has been established and is maintained by Prof. Craig C. Douglas.
©2014-2020 by www.dddas.org and Craig C. Douglas