In 1928 the passage of the McSweeney-McNary Forest research Act authorized the USDA Forest Service to conduct periodic national forest surveys. The Act called for "a determination of the present potential forest productivity of forestland" in the United States. Since this time the Forest Inventory Analysis (FIA) section of the USDA Forest Service has conducted such an inventory service across the United States. In the southeast part of the US the first forest inventory was implemented in the 1930’s
The bulk of the data that we used for estate modeling in Georgia we acquired from the plot-level FIA database. The sampling design of the FIA plots was a combination of systematic and random sampling. Specifically, the plots have been selected through a random selection of grid points, and site productivity is described in terms of site class and site index, based on arbitrarily selected site-representative dominant tree-height measurements. Measurements recorded in the FIA program were very extensive, labor intensive and covered the collection of various and detailed information that might appeal to different interest groups. Some examples of gathered records included information on dead trees and cavities, wildlife cover and habitat types, recreation information and profiles of understory vegetation.
We had access to at least 3 sources of FIA data: EWDB and WWDB databases, FIADB databases, and Master Record data.
The most popular and widely used are the EWDB and WWDB databases. They provide results from the last two periodic inventories for most of states. The databases for states often have their own set of unique characteristics. For example the database for the State of Georgia contains an additional variable, basal area, calculated from the tree level. Some limitations of the Georgia data include lack of recorded height measurements, site index variables rounded to the nearest 10 feet and the maximum site index value limited to 99 feet. It should also be kept in mind that data may be labeled with a certain year but was in fact collected over a range of years. For example, collection of data later labeled as 1997 data actually began in 1992.
Files in FIADB databases from the latest two periodic inventories were merged by conversion of EWDB and WWDB files to the same format. The structure was then extended to include space for the added variables in the new round of collected data. This new structure includes changes in the table design (7 tables instead of 3) and software for processing the data (Oracle database server) and a higher level of detail (all collected field variables are included in the database). The FIADB databases are also connected to the system for tables and maps creation called the Forest Inventory Mapmaker
Master Record data can be obtained directly from the Forest Service. Its format provides tree height measurements, but they are differently defined in different inventory cycles. MR files have a "card" structure, which means that instead of separated files with data for each level (plot, tree etc) - this format has only one file with all data in it. This format has almost no documentation.
We believe that in addition to FIA data we will be able to get also Forest industry data. This is usually very detailed data covering relatively large areas. Unfortunately they are very often confidential, provided in various formats, and collected using different procedures and designs. All these factors can make this data very difficult to apply.