In 1998, the New Jersey Department of Environmental Protection (NJDEP) contracted Aerial Information Systems, Inc., Redlands, CA., (AIS) to update the state’s existing GIS land use/land cover (LU/LC) data layer. The project was intended "to maintain and improve the quality of environmental stewardship in New Jersey" in three main ways: by updating its existing 1986 land use/land cover (LU/LC) layer to 1995/97 imagery, by increasing the level of detail in the land use data set, and by generating a new statewide impervious surface (IS) layer. The resultant updated land use/land cover layer will be an important resource for change analysis, and for supporting many environmental management initiatives.

This supplemental document describes for the user the methods and mapping criteria used by AIS to capture and compile the information used for the LU/LC update. This document is divided into six sections. Users are strongly urged to read all sections of this document, as well as those documents listed in Section 7, before using the LU/LC update data sets in any analyses.

• Section 1 includes a general description of photo interpretation methods and update mapping;

• Section 2 lists the primary and ancillary sources of information used for the data capture, including a brief history of the three primary layers used by NJ DEP to create the baseline dataset;

• Section 3 describes the LU/LC update methodology, criteria, and project issues;

• Section 4 details the specific criteria used for the LU/LC update;

• Section 5 contains the attribute definitions

• Section 6 contains the project chronology

• Section 7 contains a list of additional reference documents.

This section contains a brief description of the basic principles of photo interpretation and update mapping.

Each LU/LC feature can be recognized by its photo signature; i.e. characteristics unique to that feature. These signatures are defined by color, texture, pattern, and tonal qualities on the aerial photography. By observing the context and extent of the photo signatures associated with specific LU/LC types, the photo interpreter is able to identify and delineate the boundaries of LU/LC features.

In addition to a photo signature, the interpreter has to also understand the relationships between LU/LC types. The context in which a photo signature occurs can be an important factor in properly classifying the LU/LC feature. For example, large rectangular buildings on the photo can be either industrial or commercial land uses, making them difficult to classify without additional information. Industrial uses are often built in less desirable areas while commercial uses are usually located in areas of high traffic. If the buildings in question are located next to railroad tracks, away from well-traveled highways, then context is an important indicator for determining what is likely an industrial land use.

Collateral sources are useful to the photo interpreter as they help to identify specific LU/LC features and provide a background context against which the photo signatures can be compared. County and city street maps, USGS topographic maps, and existing vegetation maps are some of the more important sources used during the photo interpretation process. The usefulness of existing collateral sources is directly related to the detail, accuracy, and timeliness of the information they provide.

If the above sources are not sufficient for the interpreter to confidently assign a LU/LC class, on-site field visits are required. In addition to answering polygon specific questions flagged during the photo interpretation effort, on-site field surveys serve to verify correlations previously established between photo signatures and LU/LC types, ensuring that the interpretations are as error free as possible.

Update mapping is the process of re-mapping the spatial and attribute data of an existing dataset using updated sources of information for the purpose of change detection and trend analysis studies. When the attributes of the finished products are analyzed in a GIS, areas of change are exposed. Ideally, input data, mapping criteria, and methodology of the update should be the same as the original data compilation in order to make accurate comparisons.

Color Infrared (CIR) hardcopy stereoscopic photography, dated March 1995 and March 1997, was used as the primary source to interpret the LU/LC and IS for the LU/LC update. When stereo-paired, the imagery allowed the interpreter to see height and signature distinctions between vegetation classes, to note the general health and condition of the vegetation, and to identify topographic features on the landscape. In urban areas, the increased resolution provided for accurate interpretation of building heights, land use, IS percentages, and other details.

The base for the LU/LC update was the United States Geological Survey’s digital CIR orthophotos, dated March 1995 and March 1997. The digital orthophotos were compiled from the same hardcopy CIR imagery used for the photo interpretation effort. One advantage to using digital photography as the project base was the wide range of scales available, using ArcView software zoom capabilities, to view the imagery. This capability resulted in more accurate and detailed delineations of the LU/LC data.

The following is a brief description of the three major GIS layers used by NJ DEP to create the baseline dataset for the LU/LC update project. An understanding of how each of these layers was created is important as it had a direct effect on the development of mapping criteria for the LU/LC update. For more information, please refer to the project specific metadata maintained by NJ DEP.

In 1986, NJ DEP contracted Environmental Research Systems Institute (ESRI), with AIS as a subcontractor, to create a statewide environmental database for input into the department’s GIS. Four thematic layers were chosen for data capture: land use/land cover, soils, geology, and flood-prone areas. The data were generated using the Integrated Terrain Unit Mapping (ITUM) concept, pioneered by AIS. In this procedure, the individual data layers are first mapped or re-complied onto the same basemap product. Once these separate layers are created and standardized, they are integrated into one data layer, although each polygon in the final ITUM product has attribute values specific for each of the original data layers. (A full discussion of the ITUM process is given in IITUM.DOC)

The ITUM project was conducted over a period of nine years, beginning in 1986 with the Camden County pilot and ending in 1995 with the compilation of Hunterdon, Sussex, and Passaic Counties. There were a total of seven phases, each phase consisting of one or more county groups. NJ DEP personnel were involved in all decisions regarding the mapping methodology, criteria, and technical aspects of the ITUM database development.

The ITUM project LU/LC layer was mapped using photo interpretation and manual cartographic techniques. CIR stereo-paired imagery, dated 1986, was interpreted through a stereoscope while polygons were drafted onto a mylar overlay registered to a 1:24,000 black and white orthophoto basemap. The LU/LC features were classified using a modified Anderson Level II/III classification. The minimum mapping unit (MMU), i.e. the smallest area mapped as per contractual agreement, was 2.5 acres.

Concurrent with the ITUM project, NJ DEP contracted ESRI to create a statewide hydrology dataset. The hydrology data were derived from the USGS 7.5’ topographic digital line graphics (DLG) files. DLG hydrology data were compared to the orthophoto features. Where discrepancies occurred, the DLG delineations and codes were revised to fit the orthophoto image. This hydrology data set included lakes, ponds and reservoirs, as well as larger river and streams. After editing the hydrology to fit the photo basemaps, the polygon water features were integrated into the ITUM LU/LC layer.

Also concurrent with the ITUM project, NJ DEP started work on the State’s Freshwater Wetlands Mapping (FWW) program . The FWW program entailed mapping the State’s non-tidal wetlands and water bodies using the United States Fish and Wildlife Survey’s (USFS) Cowardin classification system. NJDEP’s Upper Wetlands Boundary (UWB) line coverage was used to demarcate tidal from non-tidal areas. The UWB was defined in a separate mapping program to identify areas of the state which were under tidal influence. These areas would be regulated under a specific tidal areas regulatory program. Areas above the UWB, hence not tidally influenced, were to be mapped and regulated under Freshwater Wetlands regulations.

The NJ FWW coverages were created using a number of different sources and data compilation techniques, including photo interpretation, on-site field surveys, and data from the National Resource Conservation Service’s soil surveys. The project was done for the state by MarkHurd Inc., Minneapolis, MN., who sub-contracted Greenhorne and O’Mara, Greenbelt, MD, to perform the actual photo interpretation and delineation . The FWW polygons were interpreted from the same CIR photography as the ITUM project, but mapped onto new 1:12000 scale photo basemaps. Because of the larger basemap scale, the MMU for this project was 1 acre. The delineations were then scanned, converted into ARCINFO coverages, and coded. The FWW data have since been incorporated into a number of NJ DEP reports, maps, etc., and have been used extensively as a reference data layer for wetlands permitting and wetlands regulatory programs above the UWB.

The polygons which were a part of the FWW mapping program can be identified in the LU/LC update layer by referring to the LU86, LABEL86 and TYPE86 fields for wetlands polygons (see CODELIST.DOC or ANDERSON.DOC for specific code listings), and noting if the COWARDIN and CLASS fields have entries. Only polygons mapped in the 1986 FWW program will have valid entries for these two fields. These attributes were included in the LU/LC update data layers specifically so that the FWW polygons could be identified. (See additional comments on FWW codes in Section 4, under UNIQUE CODE SITUATIONS).

After the completion of the ITUM and FWW projects, NJDEP merged the 1986 LU/LC and the FWW coverages into a single GIS dataset, which formed the baseline data set for the update project. Because of an improved basemap, larger mapping scale (1:12000 vs 1:24000), and increased field work to support the delineations, polygons delineated under the FWW program took precedence over polygons delineated under either the ITUM project or the hydrology project. (The FWW and ITUM/LULC integration is discussed in more detail in FWWINTEG.DOC).

Current city, county, and regional street maps from the following publishers, Patton, Hagstrom, ADC, Franklin, and Geographia, were used help the photo interpreters classify the land use polygons. Although stylized in their depictions, these maps were very useful for identifying major roads, parks, shopping malls, churches, and municipal buildings. Where the street map information did not correlate to the photo signature, the area was flagged for on-site investigation.

These maps were useful in land use and topographic feature determination. Specifically, they were very helpful in the identification of utility right-of-ways, transmission line substations, small mining operations, railroads, airstrips, and other features not typically shown on street maps. Although available for the entire state, the usefulness of these maps was somewhat limited by the relative age of the information shown on them.

The LU/LC update effort was conducted using computer interactive photo interpretive techniques and ArcView software. The baseline LU/LC dataset was a compilation of the 1986 LU/LC, FWW, and statewide hydrology layers maintained by the State. NJ DEP staff prepared the baseline datasets, as ArcInfo coverages, and delivered them to AIS. These were coverted to AV shapefiles for editing.

Where existing NJDEP data layers had been developed by county tiles, updated layers were to be created by Watershed Management Areas (WMA). WMA boundaries are based on actual physical drainage basin boundaries, and, therefore, define more realistic environmental management regions than do political boundaries. Prior to the start of this project, the state identified 20 Watershed Management Areas. The baseline data layers delivered to AIS were based on these 20 original WMA tiles. NJ DEP prioritized the mapping order for each WMA, establishing eight separate mapping phases over the project’s two-year timeframe. This schedule is included in Section 6.

Each WMA data layer received by AIS was further subdivided into 1:12,000 scale quarter-quad modules based on the USGS topographic quadrangle. Digital CIR orthophotos, dated 1995 and 1997, which are also based on the USGS quarter-quad tile, were used as the update project base imagery. Corresponding hardcopy CIR stereo photos were compared to the 1986 hardcopy CIR stereo imagery, and hardcopy basemaps to detect areas of change. The classification system used was a modified Anderson et al., level III/IV classification that provided the parameters for proper and consistent coding of the LU/LC feature classes and subclasses. (Specific codes used are given in CODESLIST.DOC, with a full description of the categories given in ANDERSON.DOC). Impervious surfaces were also assigned to each LU/LC polygon in 5% increments based on the percentage of paved and/or built-up area within the polygon.

Because of the resolution of the 1995/97 digital imagery, a Minimum Mapping Unit (MMU) of 1.0 acre was established for this update project. This represents a significant reduction from the 2.5 acres MMU used in the 1986 ITUM land use layer. (Note, however, that the FWW polygons included in the 1986 baseline data set were mapped at 1:12000 scale, with a 1 acre MMU).

Major factors defining the improved quality of the LU/LC update over the existing baseline layer included the increased spatial resolution and accuracy of the base imagery and basemaps available, and the technologically advanced data compilation techniques that were used. The ITUM project used basemaps that were not true orthophotos, while those created from the 1995/97 images were true ortho-rectifications. Also for the ITUM project, drafting was done manually, at a scale of 1:24,000, onto mylar overlays registered to the hardcopy black and white images. The LU/LC update, on the other hand, utilized CIR digitally rectified orthophoto imagery as the project base. Polygon delineations and polygon editing were done digitally, using specially modified ArcView tools, at scales normally between 1:2,000 and 1:5,000. These allowed for more accurate feature line placements, and more detailed land use/land cover determinations at a smaller mmu than previously possible.

While the improved imagery and data capture techniques have resulted in a more detailed and accurate final data set, special problems were created during the update process because of the differences in basemap quality. Even in areas in which no change had occurred, polygons in the 1986 LU/LC data layer, often did not align exactly with the same feature on the 1995/97 digital imagery, with the mis-registration due solely to the difference in basemap accuracies.

Contractually, AIS was not bound to correct registration problems arising from the different sets of base imagery. However, without accounting for differences due to mis-registration, true land use/land cover changes would not be captured as accurately as possible. Therefore, AIS opted to go beyond the contracted scope of work and performed limited registration adjustments to the original land use polygons as one of three "retroactive mapping" steps of the baseline data layer for the update, thus creating a more useful tool for the State’s GIS analysis needs. These retroactive registration adjustments were not performed on FWW polygons per NJ DEP’s mapping criteria (see Section 4 of this document). The other two retroactive mapping steps involved re-mapping areas that did not undergo change in order to meet the more detailed classification, and to meet the new MMU. These steps are discussed in more detail in Section 4.

One additional issue that the user should be aware of while using the LU/LC update dataset is that the data represents the LU/LC as it existed at the time of the CIR photography, 1995 and 1997. The photo interpretations for the update were not started until the fall of 1998 and the last on-site field verification survey was completed in August 2000. If on-site field surveys revealed that the land use had undergone change from the 1995-97 photography, the polygon was coded for the 1995-97 photo signature, not the 1998-2000 land use. Therefore, although the CIR photography is relatively current, the information available for land use and land cover interpretation will be a minimum of 3 – 5 years old at the time of project completion.

The following outlines the basic procedures used for the update mapping effort. These are general in nature and grew out of the unique characteristics of the project. Specific criteria of the project are discussed later in Section 4.

1. Registration – the ITUM project LU/LC delineations were corrected for nearly every existing polygon, with the exception of FWW polgyons.
2. Comparison – concurrent with the registration, the 1986 photography and 1995/97 photography were compared for feature changes.
3. Delineation – Areas of change were added to existing polygons.
4. Classification – Each polygon was coded as changed or not changed with the appropriate Anderson level III or IV category, in accordance with the criteria discussed in ANDERSON.DOC.
5. Impervious Surfaces – Each LU/LC feature was assigned an IS percentage, from 0 – 100, based upon the estimate of experienced photo interpreters. Land use/IS relationships previously established by the National Resource Conservation Service were used as baseline values for the built-up land use categories.
6. Edgematching – The 1:12,000 modules, when completed, were edgematched to adjacent modules for consistency of polygon delineations and coding. As all of the modules within a WMA were completed, the WMAs were edgematched to each other. The result is a continuous, unbroken state-wide dataset.

Upon completion of these six steps, a preliminary ArcView 3.1 shapefile of the data was delivered to NJDEP for review. If necessary, revisions were made to the dataset per NJ DEP comments. The shapefile was considered final when all of the adjacent WMAs had been mapped and edgematched across WMA boundaries, and all necessary revisions had been made.

Field verification was performed in conjunction with the photo interpretation to ensure that the data produced was as error free as possible. Eight on-site field surveys were conducted over the course of the two-year project to view areas with ambiguous photo signatures, and to ensure the accuracy of correlation between interpreted LU/LC categories and photo signatures. Each survey lasted seven working days and consisted of a two-person team driving pre-determined routes to visit polygons flagged for on-site review.

Preparation for each field visit was initiated during the photo interpretation step. At that time, all ambiguous photo signatures were flagged for field review. Upon completion of the interpretations for the field survey area, color plots containing the LU/LC polygons, codes, and imagery were created. The flagged polygons were highlighted in a different color for easy identification while in the field. The driving routes and the number of polygons visited per day were established to ensure the most efficient use of the limited survey time.

In the field, the two-person team visited or attempted to visit each of the flagged polygons. In addition, they surveyed other polygons en route to verify the accuracy of the interpretations. All notations were made directly to the field plot. All field revisions and comments were entered directly into the shapefile database upon the field team’s return to the office. Field observations of over 40,000 polygons were made during the eight field survey efforts. The chronology of the field visits is included in Section 6.

This section describes the final criteria as developed by AIS and NJ DEP. Due to changes in the data capture methodology, LU/LC classification, and minimum mapping unit from the ITUM LU/LC mapping project, several new and modified criteria were implemented for the LU/LC Update. Many of these guidelines were developed over the course of the project in response to new situations or to modify existing situations. As the procedures evolved they were submitted to NJ DEP for approval. Any discrepancies in mapping procedure from the beginning of this project to the end were due to this evolution, and were recognized by both AIS and NJ DEP.

These criteria supersede all previous criteria for this project

One of the biggest issues on the LU/LC Update was how to deal with the incorrect registration of the baseline data to the 1995/97 orthophoto base (as discussed in Section 3). Contractually, no registration correction was to be done in areas where physical LU/LC change had not occurred. However, the final product would have been far less successful as a trend analysis tool if no registration corrections were made. AIS chose to make the changes and as the LU/LC update progressed, almost every non wetland polygon line was adjusted to correct the registration of the baseline data. Criteria were developed to deal with areas of change occurring in the same locations as the ubiquitous linework registration.

• Linework between LU/LC features was not adjusted when a change to either one or both resulted in both units sharing the same LU/LC and IS codes. This occurred in instances of physical change to the LU/LC feature within unchanging boundaries, and/or actual changes to the dimensions of the LU/LC feature. This was known as the "Same Code" rule.

In all other instances between LU/LC features, whether change is involved or not, linework was adjusted to correct registration when needed.

As a separate coverage included in this LU/LC update, wetlands were afforded special consideration with regard to mapping change.

Non-tidal wetland polygons in this project were derived from New Jersey’s Freshwater Wetlands (FWW) mapping program. The original FWW coverages were mapped to an accurate 1:12000 true ortho-photo basemap, and they are currently in active use at the state level for permitting, development, and other environmental uses. Therefore modifications to the existing FWW delineations were carefully restricted so as not to affect the validity of existing programs utilizing the data.

FWW wetland polygon boundaries were not altered under any circumstances. Some coding changes were performed, as described below, but all 1986 linework remained unaltered.

• When an unchanged FWW wetland incorrectly included existing, unchanged built-up or paved areas, these built-up areas were coded out of the wetland polygon if they could be aggregated with the adjacent LU/LC class.

• When an unchanged FWW wetland incorrectly included existing, unchanged features that were not built-up or paved, nothing was done. Existing discrepancies in wetlands where the IS equaled zero were not altered.

• If a natural wetland area appeared to be significantly different than the ITUM project LU/LC code given but had not changed, the Cowardin code was reviewed for possible answers:

• If the Cowardin code accounted for the unusual LU/LC code, nothing more was done. For example, if the photo showed shrubs for a polygon that was coded as 6240 (Emergent), and the Cowardin code read "EM/SS" (Emergent and Shrubs) the shrub classification would have accounted for the shrubs as seen on the photos.

• If the Cowardin code did not account for the difference, the polygon was flagged for the attention of the NJDEP and a comment was added about what the correct code or land cover appeared to be.

• If a natural wetland area appeared to be significantly different than the ITUM project code given, and had changed:

• If there has been a "positive" change to the vegetation (growth), the wetland code was adjusted to whatever was appropriate and was coded as a change. This often happened in areas where trees grew from emergent or shrub categories, or when coniferous shrubs overgrew deciduous shrubs.

• If there has been a "negative" change to the vegetation (i.e. it had been cut back, removed, died off, or disturbed in some way) with no built-up land use, the correct code was 7430 (Disturbed Wetland).

• If a wetland underwent a change to a specific non built-up urban or agricultural land use, it was coded as the wetland equivalent of that land use.

• If the change resulted in an area of maintained grassy lawn, it was coded as 1750.

• If the change resulted in an athletic field, golf course, ball field, or other recreational land use (but not built-up or filled), it was coded as 1850.

• If the change resulted in an agricultural land use, it was coded as 2140.

• If a wetland had been cleared for a transmission line or other utility right-of-way, it was coded as 1461.

• If a wetland underwent a change and had been built up, the changed area (including both built-up and non built-up areas of change) was delineated and coded for whatever non-wetland land use and IS codes were appropriate. If the change took place throughout the entire wetland polygon, then the entire polygon was coded as a change. This was also governed by the criteria of MMU mapping.

• If a wetland had been filled it was coded with the appropriate non-wetland LU/LC code.

• All ITUM project codes of 8000 in which no change had occurred were reclassified to 1750 if non-recreation use, or 1850 if recreational use (ball field, golf course, athletic field).

• If a 2140 agricultural wetland was no longer being cultivated and was not built-up, maintained lawn, or a recreational use, then it was coded as 2150 (Former Agricultural Wetland).

• If a mappable major highway crossed over a FWW water body or wetland, the highway was delineated across the water or wetland so that the highway remained continuous. If the highway bridge existed previously, the polygon was coded to match the highway, and noted for NJDEP review

• A 7430 that changed naturally and significantly was changed to the appropriate wetland category.

The FWW mapping program did not, as a rule, delineate areas which were tidally influenced, so tidal wetlands were delineated from the 1:24,000 photo basemaps as part of the ITUM LU/LC layer. In addition, tidal water features, including tidal rivers, bays, tidal ponds, etc., were not delineated in the FWW program, so the origin of these features was the 1:24000 USGS digital line graphs for hydrology. The resulting polygons did not include a Cowardin classification. In addition, tidal wetlands polygons were, in part, defined by delineations of the hydrology layer integrated into the ITUM LU/LC data set. Furthermore, scale limitations and relative age of these USGS maps significantly affected registration of many of the polygons. Therefore, codes and boundaries were altered as necessary to register them to the photos and to update them.

If a tidal water body polygon was approximately the same size and shape as the water body on the imagery, but the linework was off registration by 60 feet or less, then it was usually not adjusted. In instances where the offset was greater than 60 feet, or when it detracted from the overall quality of the data (i.e. in built-up areas) then revised delineations were made, and noted for NJDEP review

.

The following details the delineation criteria used in the update project. Although the LU/LC Update was conducted on-screen using heads-up digitizing, many of the delineation standards established during the manually compiled ITUM mapping project were applicable to the present project.

• When a road separated differing LU/LC features, the polygons for these features followed the road centerline. Although roads often include a right-of-way on either side, these slivers of land between the roads and the adjacent LU/LC features were aggregated with the LU/LC unless they could be reasonably extracted.

• Where roads and features on the orthophoto differed from those represented on other collateral sources, the location on the orthophoto was used to depict the feature.

• Boundaries are followed as closely as possible. Where boundaries were particularly jagged, however, some cartographic smoothing was deemed appropriate. This was generally for use along natural boundaries and was rarely needed along man-made boundaries.

• Linework was kept straight and used right angle corners
• Lines were placed down the centerline of boundaries (such as roads, railroads, and tree lines)
• Structures were not dissected

Relationships between land uses and impervious surfaces previously established by the National Resource Conversation Service were used as baseline values for the built-up land use categories. Values noted for each polygon, however, were based upon the percentage of impervious surfaces within a given polygon as interpreted from the digital imagery.

• IS was interpreted in 5% increments, from 0% to 100%.

• Impervious surface percentages were assigned to each LU/LC unit based on an eyeball estimate of the paved and/or built-up area within the polygon.

• Land use/impervious surface relationships previously established by the National Resource Conservation Service were used as baseline values for the built-up land use categories.

• Where a single LU/LC polygon contained areas of differing IS and each area met the one acre MMU, the polygon was subdivided and assigned different IS values.

• All built-up and/or paved land uses were mapped as impervious surfaces and assigned IS values greater than 0%.

• Hard packed earthen areas (earthen dams excepted), gravel areas, and natural rock areas (including road cuts) were mapped as pervious surfaces.

• Compared similar land use signatures as a consistency check.

• Compared similar IS-valued polygons as a consistency check.

• Non-built up polygons less than 5 acres in size that bordered a road were generally assigned a minimum IS of 5%.

• IS for all water bodies, including concrete-lined reservoirs, were 0% unless other below-MMU paved or built-up areas existed within the polygon.

• Wetlands: If the polygon was 1 acre or larger in size, the IS code was assigned based on the amount of built-up area within the polygon. If the polygon was less than 1 acre in size, the polygon was assigned the same IS value as the adjacent or surrounding polygon.

The following guidelines were compiled for below-MMU LU/LC features and polygons. They hold true for differing, below-MMU polygons existing from the ITUM project or resulting from registration and re-mapping on this update, as well as for LU/LC features that were below-MMU for both projects:

• In general, residential land uses took precedence over other built-up land uses. If a below-MMU residential land use was adjacent to a below-MMU commercial land use, the commercial area was absorbed into the residential land use.

• The FWW delineations, including below-MMU FWW polygons, was not adjusted or altered under any circumstances as per contractual agreement.

• If a below-MMU LU/LC feature was surrounded by a number of other above-MMU LU/LC features, then the below-MMU unit would be absorbed into the unit that most closely represented its class. For example, if a below-MMU residential class was next to above-MMU commercial, industrial, and forested polygons, the residential land use was absorbed into the commercial polygon.

• An isolated below-MMU LU/LC feature completely surrounded by a single LU/LC feature was absorbed into the surrounding LU/LC class.

• Some remnant polygons remaining from the ITUM project were merged according to the criteria above if they were not significant to the overall classification of the area.

Below-MMU polygons were left intact or created in the following situations:

• Where one or more of multiple, adjacent LU/LC features had changed, now shared the same LU/LC and IS coding, and together equaled an area greater than or equal to 1.0 acre. This was an aggregation of like polygons where a physical change among them prevented merging.

• Certain non-FWW land uses could exist when adjacent to their "partner" FWW wetlands, if together they equaled an area greater than or equal to 1.0 acre. This was acceptable for the following code "partners" only: 1400/1461, 1400/1750 (for catch basins only), 1700/1750, 1800/1850, 1804/1850, 2100/2140, and 7500, 7400, 7300/7430.

• Isolated below-MMU polygons completely surrounded by FWW wetlands were left as is.

• Some LU/LC features were mapped even if below-MMU and isolated, provided they were significant to the overall classification of the area. If the 1.0-acre MMU had been strictly observed, these units would not have been captured in the database. By mapping certain LU/LC classes smaller than the 1.0 acre MMU, the user is given a more accurate depiction of LU/LC conditions in these areas.

• Some remnant polygons from the ITUM project were also kept provided they were significant to the overall classification of the area.

• When an unchanged FWW wetland incorrectly included existing, unchanged built-up or paved areas, these built-up areas were coded out of the wetland polygon (an "R" in the Comment) field if they could be aggregated with the adjacent LU/LC class. If so, these areas were strictly delineated, usually resulting in below-MMU polygons.

• The digital image took precedence over the stereo pair photography, even in situations where the original photo was dated 1997 and the digital image was from a 1995 photo. While the base year for the photographic mission was 1995, some photography had to be captured in 1997. Problems with cloud cover, or other atmospheric conditions, and snow, heavy leaf cover, and other ground conditions, often necessitate re-flying an area in a second flying season to capture acceptable photography. During the orthophoto creation process, the digital image is created from one frame that covers the entire scene to be represented by the quarterquad image, with adjacent images covering portions of the quarterquad scene used to create the stereo model. Therefore, it is possible that the scene represented by the digital image was from 1995 while the photos examined to interpret some areas may be from 1997. In these cases, the digital image always took precedence since line edits were being done over the digital image file.

• When 1995 and 1997 digital images both covered the same area, the 1997 images took precedence. Adjacent digital images, which may be of different years, have areas of overlap. In these cases, features in the overlap area were delineated using the 1997 image.

If urban/built-up use occurred together with non-urban use, the urban use took precedence over the non-urban. For example, a heavily wooded residential area was coded as residential. Exceptions were:

• Within urban areas, built-up land uses took precedence over non built-up LU/LC.

• Dense trees in recreational areas or cemeteries were separated if no built-up areas were included.

• Transmission line corridors – Although an urban/built-up land use, these corridors took precedence over only vacant land and shrubs. All other LU/LC took precedence over these corridors. See the attached listing of specific criteria for additional rules on transmission lines.

• 7000 categories normally took precedence over other, non-urban, land coverages.

Due to the dynamic nature and length of this project, and to its many unique characteristics, the mapping criteria and the LU/LC classification underwent modifications over the course of the project. This has resulted in some new categories being added as the project evolved. While some preliminary data layers were developed before these codes had been standardized, every attempt was made to use these codes in preparing the final data sets. Some minor discrepancies, however, may exist with the codes listed below. The following codes were added to the modified Anderson classification at the beginning of the LU/LC update:

The following codes were added as the LU/LC update was already underway to account for specific areas or features found to have unique characteristics not already accounted for:

• For burned areas where the vegetation remaining was still relatively intact (where the vegetation and the land cover could be readily determined), the polygon in which the burned area was included was not coded differently. New polygons were not delineated around these "lightly" burned areas.

• For upland areas where a severe burn occurred (blackened, tree crowns gone, vegetation and cover indiscernible), the area was delineated separately and coded as a change with a LU/LC code of 4500.

• For wetland areas where a severe burn occurred (blackened, tree crowns gone, vegetation and cover type indiscernible), the area was delineated separately and coded as a change with a LU/LC code of 6500.

In addition, there were several coding conventions that were developed or changed to more accurately portray less obvious transition zones between codes.

For instance, a discrepancy existed in the extensive wetland marshland along the Delaware River between WMAs 17 and 18. One side of the WMA boundary was coded 6110 (saline marshes) while the other was coded 6120 (freshwater marshes) for the same marsh. The two codes could not co-exist without a physical break between them, but without the benefit of additional data it was impossible to determine where the saline/freshwater transition would actually occur. Therefore, the confluence of Alloway Creek was chosen by NJ DEP as an arbitrary point along the Delaware River to demarcate this transition. Downstream of this point was coded 6110, while upstream was coded 6120.

Another example of possible discrepancies in interpretation of different land use types involves the 1700 and 4410 code. The current Anderson definition does not provide a clear distinction between the two classes. Therefore, in general, a 1700 code was used to define areas of maintained lawn in residential or commercial areas, as well as vacant urban land. A 4410 code was used for open spaces not carefully maintained or not a vacant lot. Despite attempts to carefully define each, inconsistencies of interpretation do exist between the two, especially in transitional areas between urban and non-urban areas.

One other transitional area was the difference between the 1110 and 1120 category. Again, the Anderson classification does not provide a clear distinction between the two types within the single-family residence in the 1/8 to 1/5 acre lot size. Where the lot size of the feature fell into the "overlap" zone of the two classes, the surrounding context of the area was used to determine whether a code of 1110 or 1120 was more appropriate. In general, residential areas in urban and old downtown suburban areas were assigned a code of 1110 while the 1120 code was used in the non-downtown suburban areas. But, again, the difference between an old downtown residential area and a suburban area is subjective, therefore making it more difficult to consistently assign a LU/LC class.

Because of the nature of the FWW data set used to create the baseline LU/LC layer, its use as a reference data set for NJDEP regulatory efforts, and the scope of the LU/LC update project, values included in the COWARDIN and CLASS fields in the attribute table for the LU/LC update layers were not updated if the wetland polygon has undergone a change. Assigning new Cowardin classifications was beyond the scope of this present mapping effort. This is true whether the FWW polygon is still a non-tidal wetlands, or has changed to another land use or land cover type. If the FWW wetlands polygon has not changed, indicated by a CHANGE86 value of 0, the COWARDIN and CLASS values can be applied to the polygon both as it existed in 1986 and as it existed in 1995/97. However, if the FWW polygon has undergone any change, indicated by a CHANGE86 value of 1, the COWARDIN and CLASS values can only be applied to the polygon as it existed in 1986. New COWARDIN and CLASS values would have to be determined on a case by case basis.

(I.e., modifications or corrections to the project due to changes in criteria rather than physical changes).

• If a LU/LC feature was below-MMU in for the ITUM project but met MMU for the update, a new polygon was delineated for the feature, given the appropriate 1986 code, and noted for NJDEP review.

• If a LU/LC feature was below-MMU in the ITUM project but met MMU for the update, and there had also been a change or addition, the entire polygon was delineated and coded as a change.

• If a LU/LC feature had already been delineated for the ITUM project, and there had been an addition, the new portion was delineated and coded as a change; the old portion remained unchanged.

• Because road centerlines were used as the boundary between polygons on either side, lines were adjusted to follow roads (assuming that the road was present and unchanged for both projects). This guideline was adhered to for all polygons, whether changed or unchanged. The lone exception to this was in instances where the "Same Code" rule applied.

• Because the LU/LC update used a more detailed version of the Anderson Classification, many codes were changed simply to reflect this greater detail.. For example, 1100 was changed to 1110, 1120, 1130, etc.

• A minimum-width unit of 60 feet was used when mapping change along long, linear features:

• Changes along coastlines – and shorelines of the Delaware River – due to erosion or deposition greater than approximately 60 feet were delineated and coded as a change. Coastal changes usually resulted in LU/LC codes indicating ocean or beach. For the Delaware River, LU/LC changes were coded with wetland codes consistent with surrounding wetlands.

• The minimum width unit of 60 feet was also used when extracting transportation corridors. For consistency between projects, most transportation corridors already delineated for the ITUM project were kept, even if they did not meet the criteria of this LU/LC update.

• ***9 was a code left over from the ITUM project to flag polygons for field verification. These polygons were checked again in the field for this project as appropriate, and replaced with the correct LU/LC code.

• Precedence was placed on land cover rather than land use for purposes of IS. For instance, open spaces were coded as such, even if on the property of a commercial area.

• Some coastal low-lying areas were originally coded as 5410 when no water was ever present. If non-FWW, these polygons were changed to 6110, and noted for NJDEP review.

• Cranberry bogs – The ITUM project code for this land use (2260) was no longer a valid code for this update. The wetland portions of existing 2260s were mapped under the FWW program, and so were given an appropriate wetlands code when the FWW layer was integrated into the ITUM LU/LC layer. Remaining upland portions of polygons originally coded as cranberry bogs were coded as 2200 or 2400 depending upon the surrounding land uses.

• In general, separate, distinct agricultural fields in the process of being developed were coded as 1700 or 4410 (as appropriate) if not directly under construction. This situation could often be recognized by new houses built on a former agricultural field where there was an obvious separation between that particular field and similar, undeveloped fields nearby. The houses indicated that the field had been sold for development; therefore, even if no obvious change to the remaining portions of the field was evident, its land use had changed. In these instances, a change to the land use directly affected accurate coding of the land cover.

• In residential areas, areas of 1700 were removed from house lots if there was an obvious transition from maintained lawn to "less maintained" lawn. Otherwise, maintained lawn spaces were left included in residential.

• Narrow road corridors (of any type of road) that were isolated between wetland polygons were generally coded as 1400 (Transportation, Communication, and Utilities). This resulted in occasional isolated road polygons.

Status: Identifies whether data set is DRAFT (for internal use only); DRAFT FINAL (for general distribution as a preliminary data set) or FINAL.

Approval: Date data set was approved by NJDEP

Acres: Acreage of each polygon.

LU95: 4 digit code representing land use/land cover category as of 1995/97.

Label95: A description of the LU95 code. (for example, High Density Residential)

TYPE95: Generalized land use category represented by LU95 and LABEL95 (for example, Urban)

IS: Percent impervious surface in polygon, from 1995/97 imagery

ISACRES: Acreage of each polygon covered by impervious surfaces.

LU86: 4 digit code representing land use/land cover category as of 1995/97

Label86: A description of the LU95 code. (for example, High Density Residential)

TYPE86: Generalized land use category represented by LU95 and LABEL95 (for example, Urban)

CLASS: Numeric code used in the original FWW mapping program to identify specific wetland types

(Can be used only with polygons identified as wetlands in the 1986 baseline data, or those

identified as wetlands in the 1995 data, which have a CHANGE86 value of 0.)

COWARDIN: Alpha-numeric label identifying specific wetlands types according to the USFWS Cowardin

Classification System. (Can be used only with wetlands polygons as described for CLASS above)

This section is organized according to the map compilation dates of the Watershed Management Areas (WMAs). The LU/LC update was mapped as eight phases over a two-year period beginning in September 1998 and ending in October 2000. Listed next to each phase heading is the mapping completion date, and under each phase heading is the name and number of the WMAs mapped. The field study chronology is also included. These studies included both pre and post compilation investigations.

Phase 1 – Pilot Project January 1999

Phase 2 June 1999

Phase 3 August 1999

Phase 4 October 1999

Phase 5 January 2000

Phase 6 May 2000

Phase 7 August 2000

Phase 8 October 2000

WMA 18 Lower Delaware Tributaries

Field Survey 1 – Pilot Project December 1998

Field Survey 2 April 1999

Field Survey 3 July 1999

Field Survey 4 September 1999

Field Survey 5 December 1999

Field Survey 6 April 2000

Field Survey 7 July 2000

Field Survey 8 August 2000

The following documents include additional information concerning the LU/LC update project and the baseline data layers. Users of the LU/LC updates should review these all of these additional documents.

LUREADME.DOC General comments on the 1995/97 LU/LC update project in Word 6.0 format.

Land use data layer in Word 6.0 format