Case Study Dataset Research Articles

One-class classification for oil spill detection

SAR oil spill classification is a challenging topic, which is tackled by semi-empirical ad hoc approaches supported by very qualified experts. In all cases, the feature space is empirically defined, and two-class classification approaches are used. Although this approach allows achieving acceptable operational results, there is still room for improving both the comprehension of the physical phenomenon and the performance of classification techniques. In this paper, we propose a novel approach to oil-spill classification based on the paradigm of one-class classification. A classifier is trained using only examples of oil spills, instead of using oil spills and look-alikes, as in two-class approaches. Further, since the feature space is empirically defined, we also propose an objective technique to select the most powerful one that is suited for the oil-spill detection task at hand. Results on two case study datasets are reported to validate the proposed approach.

Input Weighted Data Granulation Using Hybrid Correlation Measures With Application to Metal Properties

This paper introduces a new data granulation algorithm using significance weights on the input space of the data set. This data granulation algorithm aims to provide a more reliable way of grouping data together by directing the data granulation to favor the most significant variables of the process under investigation. Such a data granulation algorithm assists in the elicitation of the initial rule-base of a fuzzy or neural-fuzzy model. A hybrid correlation index, called Significance Index, is introduced to rank the process variables based on the linear correlation coefficient and the partial correlation measure. The new algorithm is used to classify the process variables and subsequently model and predict mechanical properties of heat treated steel. The property under investigation is the Tensile Strength and the case study data set consists of chemical composition and microstructure measurements coupled with Tensile Strength measurements.

Geostatistical visualisation and spatial statistics for evaluation of the dispersion of epidemic highly pathogenic avian influenza subtype H5N1

The aim of this study was to evaluate a range of statistical and geostatistical methods for their usefulness in providing insights into how highly pathogenic avian influenza (HPAI) subtype H5N1 might spread through a national population of village poultry. The insights gained allow the generation of disease dispersion hypotheses. The case study data set consisted of 161 outbreaks of HPAI subtype H5N1 in village poultry reported in Romania between October 2005 and June 2006. Reports of village outbreaks (%) occurred in three waves: October-December (14%), February-March (16%), and May-June (68%). Risk mapping - based on variography and kriging - was used to visualize the evolution of the epidemic. Outbreaks first appeared in eastern and southern Romania, particularly within an area that forms part of the Danube River Delta. The largest phase of the epidemic affected villages in all parts of central, southern, and eastern Romania, but outbreaks were clustered in central Romania. Outbreaks spread in an east to west direction. By using geostatistical visualisation and spatial statistics, the evolution of the epidemic could be characterised into two parts: disease introduction, local spread, and sporadic outbreaks, and long-distance disease spread with rapid epidemic propagation. This is consistent with the hypothesis that the environment and landscape (specifically the Danube River Delta) played a critical role in the introduction and initial spread of HPAI subtype H5N1 during the autumn and winter of 2005, and that the movement of poultry might have introduced the infection into central Romania during the spring and summer of 2006. Further research focusing on the spatio-temporal interface between the two parts of the epidemic might reveal how and why it progressed from a confined, local epidemic to a large, national epidemic. Such information would assist efforts to limit the global spread of HPAI subtype H5N1.

Radon Emanometry in Uranium Exploration Using Activated Charcoal : Namibian Case Studies

SummaryThe refinement and implementation of the Radon-on- Activated-Charcoal technique (ROAC), developed in the 1970’s by the South African Atomic Energy Board (Hambleton-Jones and Smit, 1980), is discussed. The objectives were to implement a radon (222Rn) detection technique for uranium exploration which is sensitive, cost effective, and relatively rapid in its application. The refinement of the technique, referred to here as RadonX, has primarily involved improvement of counting statistics and sensitivity. Radon, emanating from buried uranium mineralization, is adsorbed onto activated charcoal contained within a cartridge, fitted into the base of an inverted cup, and buried in the ground. The technique differs from alpha-sensitive radon detection systems in that it measures the gamma radiation arising from the daughter products of the adsorbed radon, namely 214Bi and 214Pb.Two case study datasets are presented. The first is from an exploration area potentially hosting uraniferous granites, where the sand cover varies in thickness from 10 to 100m. Time domain electromagnetic (TDEM) soundings provided cover thickness and conductivity data as a guideline to porosities. The second is from an orientation survey over a known buried palaeo-channel of duricrust-hosted uranium. The results from the latter area, after a 10-day burial period, showed an improved sensitivity compared to a past alpha-detection survey, conducted with a burial period of 30 days. The technique has proven to be highly effective through both residual and transported surficial cover, with good repeatability. A depth of penetration of 80m or more under favourable porosity conditions has been achieved.

The Soil Management Assessment Framework

Erosion rates and annual soil loss tolerance (T) values in evaluations of soil management practices have served as focal points for soil quality (SQ) research and assessment programs for decades. Our objective is to enhance and extend current soil assessment efforts by presenting a framework for assessing the impact of soil management practices on soil function. The tool consists of three steps: indicator selection, indicator interpretation, and integration into an index. The tool's framework design allows researchers to continually update and refine the interpretations for many soils, climates, and land use practices. The tool was demonstrated using data from case studies in Georgia, Iowa, California, and the Pacific Northwest (WA, ID, OR). Using an expert system of decision rules as an indicator selection step successfully identified indicators for the minimum data set (MDS) in the case study data sets. In the indicator interpretation step, observed indicator data were transformed into unitless scores based on site‐specific algorithmic relationships to soil function. The scored data resulted in scientifically defensible and statistically different treatment means in the four case studies. The efficacy of the indicator interpretation step was evaluated with stepwise regressions using scored and observed indicators as independent variables and endpoint data as iterative dependent variables. Scored indicators usually had coefficients of determination (R2) that were similar or greater than those of the observed indicator values. In some cases, the R2 values for indicators and endpoint regressions were higher when examined for individual treatments rather than the entire data set. This study demonstrates significant progress toward development of a SQ assessment framework for adaptive soil resource management or monitoring that is transferable to a variety of climates, soil types, and soil management systems.

Cloud-drift and water vapor winds in the polar regions from MODIS

Wind products from geostationary satellites have been generated for over 20 years and are now used in numerical weather prediction systems. However, geostationary satellites are of limited utility poleward of the midlatitudes. This study demonstrates the feasibility of deriving high latitude tropospheric wind information from polar-orbiting satellites. The methodology employed is based on the algorithms currently used with geostationary satellites, modified for use with the Moderate-Resolution Imaging Spectroradiometer (MODIS) infrared window and water vapor bands. These bands provide wind information throughout the troposphere in both clear and cloudy conditions. The project presents some unique challenges, including the irregularity of temporal sampling, varying viewing geometries, and uncertainties in wind vector height assignment as a result of low atmospheric water vapor amounts and thin clouds. A 30-day case study dataset has been produced and is being used in model impact studies. Preliminary results are encouraging: when the MODIS winds are assimilated in the European Centre for Medium Range Weather Forecasts (ECMWF) system and the NASA Data Assimilation Office system, forecasts of the geopotential height for the Arctic, the Northern Hemisphere extratropics, and the Antarctic are improved significantly.

Meeting in the middle: the challenge of meso-level integration

Pasture land occupies extensive areas and is increasingly of interest for sustainable intensification, land use diversification, greenhouse gas emission mitigation, and bioenergy expansion. Accurate maps of pasture and other managed land covers are needed for monitoring, intercomparison, assessing potential uses, and planning. Yet, land maps can be generated from different types of classification datasets – i.e. as a land use or land cover type – as well as different sources. In this study our aim was to assess and compare land use and land cover definitions for pasture, and examine variability in the resulting pasture land classification maps. First, we conducted a review of pasture definitions in commonly used mapping databases. We then performed a case study involving Brazil, a dominant global producer of pasture-based livestock. Six geospatial databases were harmonized and compared to each other and to MODIS land cover for Brazil including the Cerrado and Amazon biomes, which are internationally recognized for their ecological value. Total pasture area estimates for Brazil ranged by a factor greater than four, from about 430,000 km2 to over 1.7 million km2. Our analysis showed high variability in pasture land maps depending on the definitions, methods and underlying datasets used to generate them. The results are illustrative of a symptomatic problem for all manage land datasets, demonstrating the need for land categories studies and geospatial data resources that fully define land terms and describe measurable management attributes. Additionally, the suitability of individual geospatial datasets for different types of land mapping must be better described and reported. These recommendations would help bring more consistency in the consideration of managed lands in research, reporting, and policy development, as demonstrated here for pasture land using six case study datasets from multiple sources.

Evolving fuzzy neural networks for supervised/unsupervised online knowledge-based learning

This paper introduces evolving fuzzy neural networks (EFuNNs) as a means for the implementation of the evolving connectionist systems (ECOS) paradigm that is aimed at building online, adaptive intelligent systems that have both their structure and functionality evolving in time. EFuNNs evolve their structure and parameter values through incremental, hybrid supervised/unsupervised, online learning. They can accommodate new input data, including new features, new classes, etc., through local element tuning. New connections and new neurons are created during the operation of the system. EFuNNs can learn spatial-temporal sequences in an adaptive way through one pass learning and automatically adapt their parameter values as they operate. Fuzzy or crisp rules can be inserted and extracted at any time of the EFuNN operation. The characteristics of EFuNNs are illustrated on several case study data sets for time series prediction and spoken word classification. Their performance is compared with traditional connectionist methods and systems. The applicability of EFuNNs as general purpose online learning machines, what concerns systems that learn from large databases, life-long learning systems, and online adaptive systems in different areas of engineering are discussed.

Application of geostatistics to evaluate partial weather station networks

Climatic data are an essential input for the determination of crop water requirements. The density and location of weather stations are the important design variables for obtaining the required degree of accuracy of weather data. The planning of weather station networks should include economic considerations, and a mixture of full and partial weather stations could be a cost-effective alternative. A ‘full’ weather station is defined here as one in which all the weather variables used in the modified Penman equation are measured, and a ‘partial’ weather station is one in which some, but not all, weather variables are measured. The accuracy of reference evapotranspiration ( Et r) estimates for sites located some distance from surrounding stations is dependent on measurement error, error of the estimation equation, and interpolation error. The interpolation error is affected by the spatial correlation structure of weather variables and method of interpolation. A case-study data set of 2 years of daily climatic data (1989–1990) from 17 stations in the states of Nebraska, Kansas, and Colorado was used to compare alternative network designs and interpolation methods. Root mean squared interpolation error (RMSIE) values were the criteria for evaluating Et r estimates and network performance. The kriging method gave the lowest RMSIE, followed by the inverse distance square method and the inverse distance method. Co-kriging improved the estimates still further. For a given level of performance, a mixture of full and partial weather stations would be more economical than full stations only.

Alternate Approaches to Integrating Longitudinal Case Studies

In dealing with a large set of longitudinal case studies, researchers often assume that they can and should seek a single set of variables and explanatory theory to deal similarly with all the cases. This paper describes three approaches to the problems of cross-case integration and comparison. It also describes a set of comparative methods that can be used to diagnose case-study data-sets, to see which integrative approach is best.

Automated approaches for displaying spatially oriented time‐series data through image processing techniques

Abstract This study demonstrated the cartographic implications of automated image processing and computer graphics for the study of time‐series data. Automated statistical and image processing techniques were applied to a case study data set consisting of weekly Crop Moisture Index (CMI) values summarized at 174 state cooperative weather stations within Oklahoma for the time period between February and October, 1980. Computer generated isoline maps of the CMI values were interpolated and rescaled into a series of 32 grid matrices for input into a raster‐based ERDAS image processing software system. Principal Components Analysis (PCA) was used to develop graphic models that synthesized the multi‐temporal data into statistical dimensions that represented the most significant elements of CMI variability. Graphic models of the PCA statistical vectors were displayed individually, in conjunction with eigenvector loadings, and as composite images. Resultant images were analyzed statistically and graphically thro...

Dam‐Breach Flood Wave Models

Significant advances in the state‐of‐the art of dam‐breach flood wave modeling have occurred during the past decade. A comparative evaluation of several leading models representative of the current state‐of‐the‐art included a comprehensive literature survey and application of the selected models to several case study data sets. A dynamic routing model should be used whenever obtaining a maximum practical level of accuracy is required and adequate manpower, time, and computer resources are available. The National Weather Service (NWS) Dam‐Break Hood Forecasting Model (DAMBRK) is the optimal choice of model for most practical applications. Some applications require the capability to perform an analysis as expeditiously as possible. The NWS Simplified Dam‐Break Flood Forecasting Model (SMPDBK) is the optimal choice of model for most of these types of applications.