The 1995/97updated land use/land cover (LU/LC) and impervious surface (IS) layer was mapped for the purpose of providing trend analysis data throughout the state. All work was performed by Aerial Information Systems, Inc., (AIS), Redlands, CA., under the direction of the New Jersey Department of Environmental Protection (NJDEP).

Trend analysis studies rely upon the comparison of data from one year to another in order to identify areas of change. Under ideal circumstances, all of the data sets have been generated using the same classification, minimum mapping unit (mmu), input scale, mapping method, and mapping medium and it’s only the year of the photography that is different. Therefore, when the update is completed, the user only needs to compare the original data to the updated data set and look for new or different polygon boundaries and/or codes to identify areas of change, without having changes due primarily to different mapping methodologies affect the analyses.

Due to changes in NJ DEP environmental planning concerns, new computer technology, and base imagery quality, the compilation of the 1995/97 LU/LC update was conducted on a more detailed level than the baseline 1986 LU/LC layer, using slightly different mapping methodologies. At the same time, there were also significant restrictions placed upon the adjustment of Freshwater Wetlands (FWW). The user needs to be aware of these differences and their significance while using the 1995/97 LU/LC layer as a trend analysis tool.

To eliminate the need for users to independently compare 1986 and 1995/97 data sets, this new data set was produced by adding appropriate polygon boundaries and land use codes defining LU/LC in 1995/97 directly to the 1986 base data layer. All of the 1986 polygon boundaries and codes were retained in the updated layer, with only additions of new line work and codes to describe the 1995/97 conditions added. The LU/LC as interpreted from the 1995/97 imagery are defined in new 1995 land use attribute fields. Users simply need to compare the 1986 and 1995/97 polygon boundaries and LU/LC codes in this one updated layer to examine land use changes in any area.

The basic process for completing the update was to examine the 1986 base data layer over the 1995/97 digital imagery, in a special ArcView editing application developed by AIS, and adding the appropriate line work and codes to define the 1995/97 landscape. NJ DEP provided the baseline LU/LC layer from their GIS database. The baseline data layer was composed of two separate layers – the LU/LC from the Integrated Terrain Unit Mapping (ITUM) project and the Freshwater Wetlands (FWW) data. NJ DEP integrated the two themes into one coverage in 1997. All of the detail of the polygonal FWW maps, the delineations and Cowardin codes, was retained in the integrated product. (The single line freshwater wetland features were not included in the original integration). Also included in the baseline LU/LC data set, as part of the original land use layer from the ITUM mapping, were the statewide hydrology boundaries for areas below the Upper Wetlands Boundary (UWB) line. The specific process used to perform the integration of the ITUM LU/LC and FWW coverages into one theme, including a discussion of the significance of the UWB in the process, is discussed in detail in the FWWINTEG.doc file maintained by NJ DEP.

So that the final updated layer could be used to accurately analyze the 1986 to 1995/97 land use changes, modifications to the base data layers were undertaken during the update process to reduce the effects of having different mapping methodologies used in the 1986 and 1995/97 projects. Changes to the baseline 1986 LU/LC layer can be broken down into two categories (1) the revision of polygon delineations and codes due to the mis-alignment of the baseline data to the features in the 1995/97 digital image; and (2) the revision of polygon delineations and codes due to changes in the classification and mmu between 1986 and 1995/97.

Mis-alignment of the baseline data to the digital image was predominantly caused by the differences between the 1986 basemap and the 1995/97 orthophoto digital image creation process. The 1986 photo base maps were created as individual maps based on the USGS topoquad corner tics. The 1995/97 digital imagery was created using a sophisticated, integrated, aero-triangulation method, as a standard USGS. The integrated aero-triangulation results in much greater positional accuracy in the 1995/97 imagery than in the 1986 base maps.

When the 1986 baseline data set was overlain onto the 1995/97 digital images, there were degrees of mis-alignment between the digital image and the 1986 polygon delineations due to the relative quality of the two basemap sources. Since the mis-aligned polygons may lead to incorrect land use change analysis, many of the 1986 polygon boundaries were adjusted to fit the more accurate 1995/97 imagery prior to delineating and coding land use changes. Contractually, AIS was responsible for only correcting the registration of non-change polygons directly adjacent to polygons that had changed. In actuality, AIS corrected the registration of virtually all non-change polygons excepting Cowardin wetland and hydrology polygons.

In other instances, the mis-alignment was due to the changes in the map compilation technique. The interactive computer photointerpretation technique ("heads-up" digitizing) used to capture the 1995/97 data allowed the photointerpreter to zoom in for more detailed and accurate delineations than was possible using the manual cartographic method employed during the 1986 ITUM LU/LC project. Polygons that aligned to image features at 1:24000 scale usually did not align under closer scrutiny. The "heads-up" digitizing technology and the use of the zoom function effectively increased the mapping scale from the contractual size of 1:12000 (a four-fold increase from the 1:24000 scale of the 1986 ITUM LU/LC) to an actual mapping scale of 1:2000 to 1:5000. In addition, the mapping medium was significantly improved, from the gray toned frosted mylar orthophoto basemap of 1986 to the digital three band color infrared 1995/97 images. As a result, many polygon boundaries were adjusted despite the absence of change from 1986.

Another significant change was the more detailed classification chosen by NJ DEP and the change in the minimum mapping unit (mmu) from 2.5 acres to 1 acre. Selected LU/LC classes were mapped to a modified Anderson level III and IV instead of the level II detail mapped in the 1986 LU/LC layer. Both of these changes resulted in many polygons in the 1986 LU/LC layer being subdivided into a number of different polygons having different land use codes without any actual change having occurred. Note that in all cases where modifications to the base layer occurred, the 1986 codes were corrected to reflect the true 1986 land use category.

Mis-registered non-change polygons were not adjusted if the polygon was a Cowardin wetland or hydrology polygon. Prior to the creation of the integrated baseline data, the FWW and hydrology layers had been incorporated into a multitude of state mapping programs. Updating only those areas which had actually changed would insure that unaltered FWW and hydrology polygons would remain consistent with other mapped data incorporating FWW and hydrology polygons.

Note also that if AIS did not agree with the interpretation of the FWW wetland boundary and/or code, they identified the polygon for NJ DEP review, but did not change the delineation or code. In addition, any new wetland areas identified by AIS during the 1995/97 update were interpreted to the more generalized Anderson level only. NJ DEP may review these polygons and assign a Cowardin code if necessary.

Both the 1986 ITUM maps and the FWW maps were produced by county tile, which was a standard NJDEP mapping unit. The NJDEP has since adopted a watershed based management approach, with the tile being a watershed management area (WMA). These WMA’s are based on hydrologic drainage boundaries (HUC8, HUC11 and HUC14) developed by the USGS. There are 21 defined WMA’s in the state, 20 covering the land mass of the state with the 21^st being an entirely offshore tile comprising a three mile wide strip of the Atlantic Ocean. This last WMA was created for offshore water quality monitoring purposes only, and is not a part of this LU/LC update.

In order that this new data set was developed by WMA, the original county data layers were re-tiled by WMA boundaries, before being delivered to the contractor for updating. This involved joining the individual county data layers together, rectifying line and code inconsistencies between adjacent counties, dissolving out internal county boundaries, and clipping out data sets based on the WMA boundaries. The 1986 base maps and imagery were used during this process to insure that line and code changes made cartographic sense. As a last process, non-boundary, non-wetland polygons less than 1 acre in area were eliminated. These polygons were below the mapping resolution of any of the original data sets, and represented artifacts created during the integration process. These re-tiled, data layers are what were sent to the contractor as the base data layers for this project. They were sent as ArcInfo coverages, which the contractor converted to shapefiles for editing.

ADDITIONAL TILING PROCEDURES:

Between the start and conclusion of the update project, the boundaries of the WMAs were modified by the state. This was necessary because changes were made in the HUC8, HUC11 and HUC14 drainage basin boundaries by the USGS, and these changes needed to be reflected in the management area boundaries. The new WMA boundaries were officially adopted on 5/1/200 . It was, therefore, necessary to re-tile the data layers to conform to the new WMA boundaries.

Once all 20 1995 updated shapefiles as delivered from the contractor underwent a QA/QC check by DEP, they were converted into coverages using the ARCINFO SHAPEARC command. The coverages were CLEANED using a dangle length of 5 ft and a fuzzy tolerance of .01 feet.

The cleaned versions were then processed with the REGIONPOLY command to create the .PAT with the correct attribute information.

With the coverages created, the ELIMINATE command was used, to remove all non-edge polygons less than 150 sq. ft. Some small internal polygons were occasionally created in the update process, and since these were artifacts of the processing, they were removed.

After the artifact polygons were removed, the 20 WMA data layers were APPENDED together in a statewide coverage, and cleaned again with a fuzzy tolerance of .01 ft.

The DISSOLVE command was then used on this statewide coverage to create a seamless statewide data layer.

To insure that the new WMA data layers edgematched correctly, a statewide WMA coverage having the new WMA boundaries was UNIONed with the seamless statewide land use data layer. The twenty data layers with the new WMA boundaries could then be created using the RESELECT command, by reselecting out all polygons based on the WMA number.

Because the new WMA boundaries were often larger than the original WMA boundaries, several newly created WMA layers had to be edited in ARCEDIT to correct minor problems related to the creation of new polygons along the exterior WMA boundaries. In most cases, only the attributes of the new polygons had to be updated. In other cases, some line editing was involved. This occurred primarily in the WMAs along Delaware Bay, the Atlantic Ocean, and Raritan Bay. The reason line editing was involved was because of the inclusion of new open water areas in these three zones in the new WMA boundaries. So that these open water areas could be identified separately from open water areas included in the original delineation, it was felt that these new water areas should be isolated and given separate codes. Some line edits were, therefore, required to divide these new water areas from the original data set.

Once the line and code editing were completed, all data layers underwent analysis to check for labelerrors, nodeerrors and illegal codes. Any problems found were corrected, with new checks run to insure the changes were made.

The final step in the procedure was to convert the final coverages to shapefiles.

This was done using an AML running in Arc 8.0. All shapefiles were checked against the original coverages for polygon counts and attribute frequencies.

The following documents contain additional information concerning the LU/LC update project. Users of the data should review all documents before using the LU/LC update layers in any analyses.

UPDATE.HTML Detailed discussion of the 1995/97 LU/LC update project

ITUM.HTML Description of the ITUM methodology

FWWINTEG.HTML Description of process used to integrate FWW data sets in the ITUM

Land use data layer

FWW.HTML Wetlands Cowardin alphanumeric classification system.

CODELIST.HTML Land use codes used in the New Jersey LU/LC Update mapping project

ANDERSON.HTML Land use/land cover 4-digit code descriptions