MTLAKE was never used directly in predicting vertebrate species distributions; rather, it was a primary input used to create a number of grids that were actually used in the raster-based modeling process (see final Process_Step).
The following metadata elements are required by GAP, but do not parse using the FGDC ms parser (although similar elements can be found later in this document for several of these). For the convenience of GAP users, these elements are listed here. Data Set Identity: MTLAKE; Raster File Format: NA; Raster File Sensor: NA; Vector File Format: ARCE7; Nonspatial File Format: NA; Source Distance Resolution: Unknown; Raster File Number of Bytes per Pixel: NA; Native Data Structure: Vector.
From MTHYGAP, select each possible lake type and PUT to its own coverage: SELECT MAJOR3 = 421 and MINOR3 ne 610 for perennial lake coverage; SELECT MAJOR3 = 421 and MINOR3 = 610 for intermittent lake coverage; SELECT MAJOR3 = 111 for swamp/marsh coverage. (Codes in MAJOR3 and MINOR3 were evaluated and simplified for use in MT-GAP; see MTHYGAP.TXT for details.) Select islands (MAJOR3 = 200) and PUT to a separate coverage. (Note that all islands are selected here, whether they fall in lakes or in rivers; it was not possible to reliably separate islands by type. Only those islands falling within lakes, however, are counted toward item ISLAND (see below) and included in the final coverage.) Then add attribute "TYPE" to each of the four resultant coverages and populate TYPE with appropriate lake code: 421 for perennial lakes; 610 for intermittent lakes; 111 for swamps/marshes, 999 for islands. BUILD each coverage to polygon topology.
STEP 1. Create a vector layer of the desired features, using the RESELECT command in ARC to extract the desired set from MTLAKE into its own coverage based on the TYPE, ISLAND, and various size (e.g., GE5HA) attributes. For example, to identify perennial lakes greater than or equal to 10 ha that contain islands, reselect TYPE = 421 and ISLAND = 1 and GE10HA = 1. Often, the BUFFER command then was used to identify distances to the selected features. For example, to buffer a coverage of perennial lakes by 1000 m, "buffer perenlake pbuf1km # # 1000 # poly".
STEP 2. Enter the GRID module. Setmask and setwindow to ensure that the output grids are "snapped" to the state boundary grid with pixel increment, rows, and columns matching all other vertebrate modeling inputs. Specifically, the commands "setmask stategrid; setwindow stategrid mtgapveg" were used.
STEP 3. Use a CON(POLYGRID()) statement, then a CON(ISNULL) statement, to create the desired output grid (which, for modeling purposes, must contain values of 1 and 0, without any NODATA). For example, the commands "pertemp = con((polygrid(perenlake, type, #, #, 90)) == 421, 1, 0)" and "perlakegrid = con(isnull(pertemp), 0, pertemp)" would create a 90 m grid of perennial lakes where all cells within the state are coded 1 or 0.
Numerous grids were created following this sequence.