Band Name: Map of 95% prediction intervals

Title:  Biomass map using Bayesian Model Averaging & Linear Regression

Description:

Contain errors for predicted aboveground biomass values over Anne Arundel and Howard 
Counties. 

Project:  CMS

Author:  Anu Swatantran, Laura Duncanson    E-mail:  aswatan@umd.edu

Data file created on:  2011/11/25

Content Type:   TIFF    Pixel Size:  30

Location/Study Area:  Maryland (Anne Arundel and Howard Counties)

Coordinates:   West Bounding Longitude: 310946  East Bounding Longitude: 380966

               North Bounding Latitude: 4360060.363   South Bounding Latitude: 4285900.363

Data Acquisition Dates/Temporal Coverage:  Unknown

NO DATA value: 0

Coordinate System: UTM    Time Zone: 18   Datum: NAD83

Data Format: Double Precision

Data Units: Mg/ha

Data Scaled:   No

Data Creation Method:

This dataset contains errors predicted for Anne Arundel and Howard 
Counties using the Bayesian Model Averaging + Linear Regression approach. 
The dataset was divided into five NLCD 
strata viz. BROAD, CONIFER,MIXED,WETLANDS and NON-FOREST. For each strata 
Bayesian Model Averaging was used to identify the best variables for each strata. These 
variables were then used in a linear regression model to predict biomass.
Prediction intervals were then generated for biomass values in each 30m grid cell.