How an energy provider has developed a single-source, province-wide, seamless and consistent parcel layer that improves internal support for geospatial applications and decision making across the company
Manitoba Hydro is a major energy provider for the Canadian province of Manitoba. In early 2010, under the direction of Manitoba Hydro”s Geospatial Data Services, a team of consultants from Martin Newby Consulting Ltd and staff from Esri Canada set out to design, develop and create a parcel base map. The project team used multiple Esri data models and worked with more than 850 existing disparate parcel datasets to create a parcel base map that now provides Manitoba Hydro with an integrated and consistent parcel feature.
Data modeling: Reviewing puzzle pieces
As a first step, the team performed a gap analysis to identify where the 850 source datasets could and could not provide the framework required to support the data model. The datasets came from the City of Winnipeg, the Province of Manitoba and Manitoba Hydro”s internal archives; data for aboriginal land and parkland came from Natural Resources of Canada.
The analysis phase identified various data issues — gaps in the spatial extent, limited original survey distances and bearings (coordinate geometry attributes), lack of true circular arcs, inadequate aggregate representations (plan boundaries), poor topology and a gamut of attribute inconsistencies. The team devised solutions to address each condition with a primary objective in mind — build the fabric using the best available data without redrafting existing plans of survey.
Tool development: Building the puzzle
Manitoba Hydro”s Geospatial Data Service steam selected Safe Software”s Feature Manipulation Engine (FME) to provide a repeatable process for integrating datasets from source providers into the parcel fabric, a feature class developed by Esri. FME has the capability of transforming multiple data sources and formats in various ways.
A parcel fabric is a dataset for the storage, maintenance and editing of parcels. It is created under a feature dataset and inherits its spatial reference from the feature dataset. Each of the 850 puzzle pieces was transformed from a source state into an individual, consistently formatted staging file geodatabase. The staging file geodatabase provides interim means to check data prior to loading into the parcel fabric. It also allows flexible workflows, such as reprocessing either single or multiple source datasets or performing a provincial refresh. Twenty unique workspaces were developed to manipulate the 850 different datasets in a modular style.
Once the staging file geodatabases are prepared, loading to the parcel fabric can be done using out-ofthe- box functionality of ArcGIS. But to enhance the loading process with quality control and automation, a customised loader was developed.
Data integration: Fitting the puzzle pieces
In total, more than 850 datasets from the four source providers were ultimately integrated into the cadastral fabric. In some cases, the data providers supplied known spatial accuracy statements, datums and projections. In other situations, spatial accuracy and georeferencing details became another puzzle to solve.
Martin Newby’s team of geomatics technicians has integrated over 400,000 plans of survey for the Province of Alberta over the past 13 years. The spatial alignment of a particular site in northeast Manitoba conflicted with the surrounding parcel data. The team used 1:60,000- scale orthoimagery to validate that this parcel dataset required a shift and rotation. A custom Adjustor FME workspace was built to address situations where adjustment algorithms are used to spatially conflate data and edge match with surrounding datasets. The source data required dramatic adjustment to better fit the surrounding spatially accurate data. CAD data from Manitoba Hydro”s internal archives presented a unique challenge. Over the course of many years, these site-specific survey drawings had been drafted using custom AutoCAD software, and it was presumed that the content structure was uniform. However, a comprehensive data review revealed that various layer structures and standards were used that would require extensive effort to normalise.
Many of Martin Newby’s custom batch programmes and interactive tools provided an efficient means to clean CAD data that had various topological and geometry issues such as undershoots, overshoots, duplicates, zero-length lines, and invalid elements. An extensive manual cleaning effort was required in situations where line work was scattered among various layers.
The finished puzzle
Within nine months, beginning with a team start-up meeting in February and ending with the first parcel fabric delivery in October 2010, Manitoba Hydro was able to see its initial objective realised. The utility successfully created a key GIS dataset — the Property Parcel Fabric — which enabled the growth of GIS applications to support internal business units. Now, Manitoba Hydro”s Geospatial Data Services team can easily and readily maintain the parcel fabric with the integration of new source datasets. The tools and workflow developed support this repeatable process. Further enhancements, such as the collection and use of ground control, will result in increased spatial accuracy. The ability to improve and tighten the horizontal spatial accuracy exists as core functionality within the cadastral fabric data model. Source datasets available at present have limited inherent ties to control, but with subsequent fieldwork and the additional mapping and integration of plans of survey, the survey control within the fabric can be strengthened and more fully populated.