Modeling Space/Time Variability with BMEGUI Prahlad Jat (1) and Marc Serre (1) (1) University of North Carolina at Chapel Hill
Agenda ► Introduction ► Mean Trend Analysis ► Space/Time Covariance Analysis
Introduction
Temporal GIS analysis process Read Data File Check Data Distribution Exploratory Data Analysis Mean Trend Analysis Covariance Analysis BME Analysis Data Field Screen Data Distribution Screen Exploratory Data Analysis Screen Mean Trend Analysis Screen Space/Time Covariance Analysis Screen BME Estimation Screen
Mean Trend Analysis Screen ► Display temporal mean trend ► Display spatial mean trend (Raw & Smooth) ► Model Parameter (Exponential Smoothing)
Space/Time Covariance Analysis Screen ► Display spatial & temporal covariance ► Plot covariance models
Mean Trend Analysis
Mean Trend Calculation ► Assume a separable additive space/time mean trend model ► Input Parameter Spatial Radius/Spatial Range Temporal Radius/Temporal Range ► Averaging the measurement at each MS (or at each time point) ► Find measurements within “Radius”, then apply exponential filter
Smoothed Mean Trend T radius S radius
Remove Mean Trend ► Removing the mean trend from data Z: Value SSM: Smoothed Spatial Mean Trend STM: Smoothed Temporal Mean Trend Z - (SSM+STM-mean(STM))
Calculate Mean Trend ► Click “Model mean trend and remove it from data” ► BMEGUI automatically calculate mean trend using the default parameters
Temporal Mean Trend ► Raw Temporal Mean Trend is shown in dotted line ► Smoothed Temporal Mean Trend in shown in solid line ► Zoom in/out
Spatial Mean Trend ► Two tabs – Spatial Mean Trend (Raw) Spatial Mean Trend (Smoothed) Spatial Mean Trend (Smoothed) ► Zoom in/out ► Point Layer File
Recalculate Mean Trend ► Input parameters and click “Recalculate Mean Trend” button Spatial Radius/Spatial Range Temporal Radius/Temporal Range Spatial Radius Temporal Radius Spatial Range Temporal Range Click Button
Space/Time Covariance Analysis
► Experimental Covariance (Red dots) ► Fit experimental covariance with covariance model (Solid Line)
Spatial/Temporal Components ► Two tabs Spatial Component Temporal Component Spatial Component Tab Temporal Component Tab
Experimental Covariance ► BMEGUI automatically calculate experimental covariance using the default lag setting ► User can modify the lag setting Change the number of lags Input user-defined lag and lag tolerance
Change number of lags ► Input the number of lags, then click “Recalculate Spatial/Temporal Covariance” button Number of lags for Spatial Cov. Number of lags for Temporal Cov. Recalculate Spatial Cov. Recalculate Temporal Cov.
Modify lag and lag tolerance ► Click “Edit Spatial/Temporal Lags…” ► Input lag and tolerance in dialog box, then click “OK” (Use “,” to separate the values)
Covariance Model ► BMEGUI supports homogeneous/stationary Space/Time random field Space/Time separable model Maximum four model structures Following covariance model ► exponential ► gaussian ► spherical ► holecos ► holesin
Set the number of model structures ► Set the number of model structures Input the number into text box (1-4) The number of tabs will change
Select covariance model ► Select model from the combo box ► Input sill and range, then click “Plot Model”
Clear covariance model ► Plot “Clear Plot” button ► Covariance model will be erased from the figures