Effect Of Data Quality Research Articles

Improving watershed decisions using run-off and yield models at different simulation scales

Water managers face the daunting task of balancing limited water resources with over-subscribed water users among competing demands. They face the additional challenge of taking water planning decisions in an uncertain environment with limited and sometimes inaccurate observed and simulated hydrological data. Within South African watersheds, spatial parameterization data for hydrological models are now available at two different basin management resolutions (termed quaternary and quinary). Currently, water management decisions in the Crocodile River watershed are often made at a more coarse resolution, which may exclude crucial insights into the data. This research has the following aims (1) to explore whether model performance is improved by parameterization using a more detailed quinary-scale watershed data and (2) to explore whether quinary-scale models reduce uncertainty in allocation or restriction decisions to provide better informed water resources management and decision outcomes. This study used the Agricultural Catchments Research Unit (ACRU) agro-hydrological watershed model, to evaluate the effects of spatial discretization at the quaternary and quinary scales on watershed hydrological response and runoff within the Crocodile River basin. Model performance was evaluated using statistical comparisons of results using traditional goodness-of-fit measures such as the coefficient of efficiency (C eff), root mean square of the error and the coefficient of determination (R 2) to compare simulated monthly flows and observed flows in six subcatchments. Traditional interpretation of these goodness-of-fit measures may be inadequate as they can be subjectively interpreted and easily influenced by the number of data points, outliers and model bias. This research utilizes a recently released model evaluation program (FITEVAL) which presents probability distributions of R 2and C eff derived by bootstrapping, graphical representation of observed and simulated stream flows, incorporates statistical significance to detect the sufficiency of the R 2and C eff and determines the presence of outliers and bias. While analyses indicate that the ACRU model performs marginally better when parameterized and calibrated at the quinary scale, the measurements at both scales show significant variability in predictions for both high and low flows that are endemic to southern African hydrology. The improved evaluation methods also allow for the analysis of data collection errors at monitoring sites and help determine the effect of data quality on adaptive water planning management decisions. Given that many water resource challenges are complex adaptive systems, these expanded performance analysis tools help provide deeper insights into matching watershed decision metrics and model-derived predictions.

The effect of data quality on model performance with application to daily evaporation estimation

The model performance is usually influenced by the quality of the data used in model construction. If the model performance is less affected by data quality, the resulting estimates will be more reliable. In this paper, the variation in model performance due to different data quality is explored in a field-scale application. Hence, two models, the proposed support vector machine (SVM) based model and the Stephen and Stewart (SS) model, are employed for daily estimation of evaporation at an experiment station. Five scenarios corresponding to different data qualities are designed to evaluate the effect of data quality on model performance. Additionally, the most effective meteorological variables influencing evaporation are obtained by a systematic input determination process. These most effective meteorological variables are used as inputs to the SVM-based model. The results show that the model performance decreases as the data quality decreases (i.e. the percentage of missing data increases). However, the estimation accuracy of SVM-based models is still better than that of the SS model. Moreover, the variation of model performance of the SVM-based model is smaller than that of the SS model. That is, the negative impact of different data quality is effectively decreased by using the SVM-based model instead of the SS model.

Effects of data quality and quantity in systems modelling: a case study

When constructing a probability distribution from incomplete and imprecise data, the effects of the quantity and the quality of the data are of serious concern in practical applications. Consider a situation when one is building a matrix of a joint probability distribution. For some events, the probabilities are available only approximately, and for the majority of the events they are not available at all. Traditionally, if the known values are exact values, this type of problem is dealt with by maximizing the Shannon entropy of the distribution while using the known values as constraints. In this case, however, the available information is approximate and represented by fuzzy numbers. A multi-objective optimization method is proposed that employs the well-known principles of maximum and minimum uncertainty. In this method, the Shannon entropy is maximized and, in addition, the known elements whose membership grades are as high as possible are searched for. The method is applied to the construction of an origin–destination (O–D) table of a transit from incomplete and imprecise data. The behaviour of the solution with respect to quantity and quality of available data is tested with sensitivity analysis using real-world data of four transit lines. This analysis reveals how changes in the quantity and quality of data affect the acceptable level of an O–D table. Furthermore, the issue of how to combine O–D tables developed on the basis of different sets of approximate values is examined using a method that minimizes the sum of the relative Shannon entropies.

Data quality and firm performance: empirical evidence from the Korean financial industry

Despite popular belief that timely and precise data are important and indispensable to good decisions and that good decisions are related to better firm performance, empirical research that examines the effect of data quality on firm performance is still scarce. How great an impact does data quality have on firm performance? This study empirically investigates the effect of firm-level data quality on firm performance in the Korean financial industry during 2008---2010. The results show that commercial banks have high-quality data, while credit unions have comparatively low-quality data. They also show that better data quality has a positive influence on sales, operating profit, and value added. Improving the level of data quality management maturity by one can increase firm performance by 33.7 % in sales, 64.4 % in operating profit, and 26.2 % in value added.

Comparing experimental design schemes in predictive food microbiology: Optimal parameter estimation of secondary models

In predictive food microbiology, full factorial designs are still more the rule than the exception, despite the huge experimental workload and cost related to this method. In this study, two simulation studies for secondary square-root-type models are performed to compare several experimental designs with respect to four criteria: (i) number of experiments, (ii) goodness-of-fit statistics with respect to the original model structure, and (iii) accuracy and (iv) uncertainty of the parameter estimates. In addition, the effect of data quality, quantified as the error related to plate count measurements, is assessed on the relation between model structure and experimental design. Full factorial, reduced full factorial, central composite, Latin-square and Box-Behnken designs are evaluated and compared to randomly selected datasets.As a guideline, a full factorial design should be preferred for rather simple model structures and a limited number of levels per environmental factor. For more complex cases, a Latin-square design is an attractive alternative as it does not require a priori model knowledge and provides relatively accurate and reliable parameter estimates while keeping the experimental efforts to a minimum.

Fidelity Performance of Three Filters for High Quality NDVI Time-series Analysis

Some noises still exist in standard MODIS NDVI product and need to be screened before further applications.An ideal filter should only remove noisy data,and keep the good data as much as possible.Most previous studies only concerned about the filter's performance in terms of handling noisy data,and neglected filter's capacity to retain the original high-quality data(fidelity).The present study analyzed three filters included in Timesat 2.3 tools,examined the effect of data quality on the performance of three filter algorithms,both qualitatively and quantitatively.Through comparative analysis,we concluded that Asymmetric Gaussians has the highest fidelity performance of all,followed by Double Logistic algorithm,and Savizky-Glolay algorithm performance worst.

The effects of data quality in local earthquake tomography: Application to the Alpine region

Despite the increase in quality and number of seismic stations in many parts of the world, accurate timing of individual arrival times remains crucial for many tomographic applications. To achieve a data set of high quality, arrival times need to be picked with high accuracy, including a proper assessment of the uncertainty of timing and phase identification, and a high level of consistency. We have investigated the effects of data quantity and quality on the solution quality in local earthquake tomography. We have compared tomographic results obtained with synthetic and real data of two very different data sets. The first data set consisted of a large set of arrival times of low precision and unknown accuracy taken from the International Seismological Centre (ISC) Bulletin for the greater Alpine region. The second high-quality data set for the same region was seven times smaller and was obtained by automated quality-weighted repicking. During a first series of inversions, synthetic data resembling the two data sets were inverted with the same amount of Gaussian distributed noise added. Subsequently, during a second series of inversions, the noise level was increased successively for ISC data to study the effect of larger Gaussian distributed error on the solution quality. Finally, the real data for both data sets were inverted. These investigations showed that, for Gaussian distributed error, a smaller data set of high quality could achieve a similar or better solution quality than a data set seven times larger but about four times lower in quality. Our results further suggest that the quality of the ISC Bulletin is degraded significantly by inconsistencies, strongly limiting the use of this large data set for local earthquake tomography studies.

COMPONENTS OF UNCERTAINTY IN SPECIES DISTRIBUTION ANALYSIS: A CASE STUDY OF THE GREAT GREY SHRIKE

Sophisticated statistical analyses are common in ecological research, particularly in species distribution modeling. The effects of sometimes arbitrary decisions during the modeling procedure on the final outcome are difficult to assess, and to date are largely unexplored. We conducted an analysis quantifying the contribution of uncertainty in each step during the model-building sequence to variation in model validity and climate change projection uncertainty. Our study system was the distribution of the Great Grey Shrike in the German federal state of Saxony. For each of four steps (data quality, collinearity method, model type, and variable selection), we ran three different options in a factorial experiment, leading to 81 different model approaches. Each was subjected to a fivefold cross-validation, measuring area under curve (AUC) to assess model quality. Next, we used three climate change scenarios times three precipitation realizations to project future distributions from each model, yielding 729 projections. Again, we analyzed which step introduced most variability (the four model-building steps plus the two scenario steps) into predicted species prevalences by the year 2050. Predicted prevalences ranged from a factor of 0.2 to a factor of 10 of present prevalence, with the majority of predictions between 1.1 and 4.2 (inter-quartile range). We found that model type and data quality dominated this analysis. In particular, artificial neural networks yielded low cross-validation robustness and gave very conservative climate change predictions. Generalized linear and additive models were very similar in quality and predictions, and superior to neural networks. Variations in scenarios and realizations had very little effect, due to the small spatial extent of the study region and its relatively small range of climatic conditions. We conclude that, for climate projections, model type and data quality were the most influential factors. Since comparison of model types has received good coverage in the ecological literature, effects of data quality should now come under more scrutiny.

Distribution, ecology and conservation status of the Blue-headed Macaw Primolius couloni

Until recently thought to be secure (i.e. Least Concern), the Blue-headed Macaw is now classified as Endangered in the IUCN Red List, based on an apparent decline and a population estimate of <2500 mature individuals. We review published and unpublished sources, collating records from 61 localities in Peru, Brazil and Bolivia, and compiling information on habitat use, seasonality, group size, demography, and population density. We find the species to be associated with disturbed habitats at one site, but a broader analysis revealed no significant associations with forest type, riverine habitats, degree of disturbance or altitude. By mapping locality records, and accounting for discontinuities, we calculate an Extent of Occurrence of 460,000 km 2. Range-wide data on encounter rates and flock sizes suggest that the species is sedentary and gregarious, with an overall population density of one mature individual per 10–50 km 2. Our figures for range size and density (both highly conservative) indicate that the global population estimate should be revised upwards to 9200–46,000 mature individuals. Balanced against an increasing threat from trade, these data argue for a reversal in status to Vulnerable, with a shift to Near Threatened possible in future. Given these recent fluctuations in conservation status, the Blue-headed Macaw provides valuable insight into the difficulties of using IUCN Red List criteria to assess poorly known taxa. Red List assessments should be based on extensive reviews where possible, and analyses using Red List data should consider effects of data quality.

The effect of data quality at the time of introduction of population-based screening on the estimate of programme impact using surrogate outcome measures

The aim of this study was to examine the availability and quality of surrogate outcome measure data to enable the evaluation of the breast screening programme. As the observed effect on mortality of a screening programme occurs many years after the introduction of screening, surrogate outcome measures offer an attractive alternative, allowing estimates of the effect to be calculated earlier. The exercise was undertaken by the Oxford Cancer Intelligence Unit and the Cancer Screening Evaluation Unit in collaboration with cancer registries in England, Scotland and Wales. The conclusion of the exercise was that, in general, the available data quality was insufficient to allow a precise estimate of the overall mortality reduction from breast screening to be made using surrogate measures. When a screening programme is started, it is vital that forward planning is undertaken to ensure that the necessary information is ready to be collected before the start of the programme. The use of surrogate measures is dependent on high-quality data in the uninvited group of women (pre-screening or from a staggered start over areas), and improving data quality over time after the start of the programme is of very limited value for a surrogate measures approach. The collection of appropriate high-quality information before the programme starts is, therefore, critical to measuring the success of the programme. Cancer registries and staff within the screening programme need to be ready to collect the appropriate data before the introduction of a screening programme. The exercise undertaken therefore has important implications where new screening programmes are being considered.

Effect of data quality on estimates of farm infectiousness trends in the UK 2001 foot-and-mouth disease epidemic

Most of the mathematical models that were developed to study the UK 2001 foot-and-mouth disease epidemic assumed that the infectiousness of infected premises was constant over their infectious periods. However, there is some controversy over whether this assumption is appropriate. Uncertainty about which farm infected which in 2001 means that the only method to determine if there were trends in farm infectiousness is the fitting of mechanistic mathematical models to the epidemic data. The parameter values that are estimated using this technique, however, may be influenced by missing and inaccurate data. In particular to the UK 2001 epidemic, this includes unreported infectives, inaccurate farm infection dates and unknown farm latent periods. Here, we show that such data degradation prevents successful determination of trends in farm infectiousness.

Investigating the effect of data quality on time domain electromagnetic discrimination

Using field data and numerical simulations we investigate the effect of data quality on time domain electromagnetic discrimination. Data quality decreases when measurements contain responses not accounted for by our mathematical modelling. This can include instrument noise, inaccurately reported position and orientation information, geologic contributions to the signal, and loss of validity of the forward modelling. Survey design is critical to data quality in order to have sufficient sampling of data anomalies, and also to ensure that each target is illuminated such that both the axial and transverse components of the polarization can be excited and measured. For dipole model based discrimination algorithms, success is contingent upon the accuracy with which the components of the polarization tensor can be estimated. Field data from different survey modes are analysed to identify noise sources and provide quantitative estimates of the noise in each survey. Inversion results show that increased noise levels lead to greater spread in recovered parameters. Monte Carlo simulations are performed in order to investigate the importance of other data quality factors. Analysis of inversion results from the simulations show that anomaly size, signal to noise ratio, positioning error, line spacing and station spacing all play a role in the spread of recovered parameters. Through the analysis of our simulation results we propose a figure of merit as a means of quantifying different data quality factors with a single number and relate this number to the accuracy with which parameters can be estimated.