Band Name:  BMA biomass map

Title:  Biomass map using Bayesian Model Averaging & Linear Regression

Description:

Contains 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 biomass values 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 also generated for biomass values in each 30m grid cell.

Refer to error analysis folder for high and low error maps from BMA

Biomass equations developed using the BMA + Linear Regression approach

Biomass broadleaf ~ 6.526*P60+31.046*D55 -30.350* D60 -33.804 
R-squared = 0.52 , RMSE = 97.76

Biomass conifers ~ -30.572* P20+185.089*P40 -149.048*P45+ 25.589*D40 -24.486*D45-17.701
R-squared = 0.71 , RMSE = 67.99

Biomass mix ~  13.290* P100+27.102*QMHA -40.947*QMHC+ 2.443*D70 -22.216  
R-squared = 0.62, RMSE = 84.3

Biomass wetlands ~   1.4843*P10+ 3.0150*D55-0.9306*Ksat_30m+16.1832 
R-squared = 0.62, RMSE = 75.8

Biomass non forest ~ 5.8444*P85+ 2.9436 
R-squared = 0.53, RMSE = 32.48


Overall fitted R-squared and RMSE