MAINTAINING A GIS: CRITICAL ISSUES
Shane A. Imwalle(1)
ABSTRACT: Now that many communities have GIS up and running, everyone is asking the same question: How are we
going to maintain our GIS? In spite of the hundreds of thousands of dollars poured into planning and developing GIS each
year, the industry has yet to settle on methods that consistently address critical issues involved in maintaining GIS data. It's
clear that GIS maintenance has traditionally been an afterthought; a look at the databases of many systems reveals designs
cluttered with features neither necessary nor feasible to maintain. A logical method for maintaining a GIS should be planned
early in the project, during the database design process, and included in the GIS implementation plan. This paper addresses six
critical issues to consider when creating a maintenance plan for a GIS: applications, cost-effectiveness, data sources, accuracy,
type of maintenance, and staffing. The author will also present guidelines for making decisions about critical issues in GIS
maintenance. Creating and implementing a logical GIS maintenance plan is crucial if a GIS is to be useful in the future.
INTRODUCTION
Many communities that have geographic information systems (GIS) up and running are asking the same question: How are we
going to maintain our GIS? In spite of the hundreds of thousands of dollars poured into planning and developing GIS each
year, the industry has yet to settle on methods that consistently address critical issues involved in maintaining GIS data. It's
clear that in many cases GIS maintenance is an afterthought; a look at these databases reveals designs cluttered with features
neither necessary nor feasible to maintain.
A logical method for maintaining a GIS should be planned early in the project, during the database design process, and
included in the GIS implementation plan. Maintenance tasks should be implemented during data conversion as soon as data is
received.
This paper addresses the issues that cities, counties and other GIS users should consider when creating a GIS maintenance
plan.
ISSUES TO BE CONSIDERED IN GIS MAINTENANCE PLANNING
The following six issues are critical to planning a GIS maintenance program:
Applications
One of the real values of a GIS is the ability to share data among a number of users. Participants must agree on the data to be
input and how it should be maintained. To do this, it's important to determine the GIS applications each participating agency
needs (for example, utility maintenance; customer billings; and infrastructure expansion; or land use planning; zoning; and parcel
mapping).
GIS applications will determine what mapping features should be included in the GIS. Participants must determine how each of
these elements will be updated over time. Some GIS databases include mapping features that change frequently but have no
clear or practical applications. Including features that change frequently (such as individual trees) in a GIS may not be
cost-effective because maintenance will be costly. Unless there are important applications for frequently changing features,
these elements should be omitted from the database design. GIS participants designing the database should choose features
that satisfy applications but do not require constant updates. The use of a mapping feature must always be worth more than the
cost of maintaining that feature. Participants should also determine how useful each mapping feature will be and how many
participants will benefit. If only one participant needs a feature for a particular application, then it may not be cost-effective to
include and maintain that feature. The input and maintenance of features to be used by several agencies, however, would likely
be cost-effective in the long run. Applications may also help participants determine what type of maintenance will be used.
Some applications such as E-911 (emergency dispatching) may require data to be updated regularly and often, while data for
other applications, such as market analysis, may require only occasional maintenance.
Cost-Effectiveness
The cost-effectiveness of a GIS maintenance program involves making budgetary decisions about the value of digitizing and
maintaining certain mapping features in a database. If a feature is not cost-justified to maintain based on applications, then it
may not be necessary to include that feature in the database design.
Building a database is the largest initial expense of a GIS program; but in the long run, database maintenance may cost more
than GIS development. Database development costs are directly related to the number and type of mapping features included,
which are determined by GIS applications and required accuracies. One way to help ensure cost-effective maintenance may
be to construct the base map using smaller scale accuracies (for example, 1=200 instead of 1=100).
Because the base map is the GIS foundation, the positional accuracy of data added depends upon the accuracy of the base
map. The base map is typically created from aerial photography and ground control, and its horizontal accuracy depends on its
map scale (the larger the scale, the more accurate the base map, and the more mapping features can be included). Because
fewer mapping features can be included on smaller scale maps, there are fewer features to maintain; and because fewer plots
need to be produced during input and maintenance, plotting costs are decreased. However, in most cases, smaller scale base
maps are less accurate than larger scale maps.
An option to traditional update tools for base mapping in a GIS is the use of digital orthophotos. These raster image,
"computer rectified" photographs that match traditional vector (line) maps one-to-one, in terms of scale, have proven to be a
reliable, accurate alternative base map foundation. Using digital orthophotos can be a less expensive and more attractive
option for building a GIS land base. Now being touted as a cost-effective way to maintain a GIS land base, digital orthophotos
offer capabilities such as accuracy of scale, heads-up digitizing of new and changed features, and immediate feature recognition
(because of their photographic qualities).
In planning a GIS, the developers must be certain that accuracies of the intended map scale satisfy the application demands.
However, they must also be aware that substantial cost differences exist for different levels of accuracy and output scales of
mapping.
Maps with 1=100 scale accuracies may cost two to three times more than maps with accuracies of 1=200. This is because
more aerial photographs are required to produce maps at larger scales and in turn, there are more output sheets. For example,
approximately 96 standard map sheets are needed to map a 20-square-mile city at 1=100 scale, while only about 24 sheets
are needed at 1=200 scale.
While each of the six issues has an impact on every other issue, applications and cost-effectiveness generally have the most
influence on decisions made during the development of a GIS maintenance program.
Data Sources
Municipalities and utility companies have many sources of data for updating a GIS, ranging from as-built drawings to site plans
to aerial photographs. To establish an efficient, cost-effective maintenance process, GIS decision makers should establish a
maintenance system and set policies to govern the use of these data sources for maintaining the GIS.
These policies should cover:
A GIS should have safeguards built in that determine who can and cannot have full access to the GIS data. If separate
agencies or users will continue with responsibility for their own sections of the GIS, the developers must consider permitting
some to have read-only access to GIS data other than their own. In either case--establishing a GIS maintenance department
or continuing individual update responsibilities--the key is to set and follow maintenance policies.
Accuracy
Accuracy in GIS maintenance encompasses the accuracy of data input and data maintained. During database design, GIS
participants must develop and agree on accuracy standards that will govern data input and maintenance. Accuracy standards
are determined based on GIS applications. For instance, the applications of the GIS drived the accuracy levels in terms of
positional locations for base features, types of structures for the utility system, and non-graphic (or tabular) attribution for the
parcel or utility line.
When considering digital orthophotos for GIS land base maintenance, the user will have a photographic tool to update obvious
changes, such as new structures, additional pavements, or a new landscape plan. The user becomes the photo-interpreter
whose basic tool is a photographic image with all displacements removed. Will the user attain the same level of accuracy
placing a new feature in the GIS database? Probably not. The digitizer works with only one photo image, while a stereoplotter
operator has the advantage of using a stereo model with the ability to see both sides of a building. The user is able to digitize
the location of the base of a new power pole or the center of a manhole or the edges of a new sidewalk, but is not as likely to
obtain an accurate outline of a building.
As participating agencies decide which data sources to use for maintenance, they must consider what accuracies the
applications require, which data sources provide those accuracies, and the cost-effectiveness of accessing and using these
sources. The greater the accuracy requirements, the more costly data conversion and maintenance will be. If participants want
high accuracy standards, they must decide whether such a high level of accuracy is necessary for the intended applications. If
high accuracies are necessary, participants must then determine whether the data can be maintained from sources with the
same accuracy level. It may be cost-effective to compile certain features from aerial photography during data input but not
during maintenance.
Because different applications often have different accuracy requirements, participants may find it difficult to decide on GIS
accuracy standards. One agency's applications may require data with greater accuracies than required by another agency, such
as engineering design versus community planning. Accuracy decisions made during GIS planning will likely require
compromise. Ultimately, to keep the project cost-effective, GIS participants must decide which applications absolutely
demand greater accuracy.
Some people believe that a less-accurate GIS can cost-effectively "become" more accurate through maintenance efforts. This
approach is not cost-effective because it requires a significant amount of time and effort, resulting in varying levels of accuracy
for different types of data. The most cost-effective approach is inputting accurate data from the beginning.
If participants are considering maintaining data at a lower accuracy level than that used during data input, they should
determine whether a lower accuracy will adversely affect applications. If participants decide to maintain data at a lower
accuracy level than that used during data input, they should identify these updates in some way (for example, with unique
symbology) so that users know the accuracy of updates is different from the original data input.
Type of Maintenance
Decision makers must also determine how to maintain each feature and when to initiate maintenance for that feature. Ideally,
GIS maintenance should accomplish the following goals:
Two types of GIS maintenance procedures are transactional maintenance and periodic maintenance. Which type of
maintenance is best depends on previously mentioned factors such as applications, cost-effectiveness, and the accuracy and
accessibility of data sources.
Transactional Maintenance
Transactional maintenance involves making frequent updates to the GIS as soon as the appropriate data sources are available.
The accessibility of these data sources determines whether transactional maintenance is suitable for a particular GIS feature.
Suitable data sources for transactional maintenance usually cover only small geographic areas (subdivisions or development
projects) or just a few mapping features (hydrants, utility poles). As-built drawings are ideal for transactional maintenance
because they are accurate enough for direct input of updated features. Examples of transactional maintenance are adding to the
base map from a digital orthophoto image or a subdivision plat drawing, or updating a valve record file based on the results of
processing a work order.
Periodic Maintenance
Periodic maintenance involves making changes to the entire GIS or to defined parts of the GIS only during scheduled
maintenance events. These changes, which occur over a relatively long time, are input during a comprehensive update effort.
Ideal data sources for periodic maintenance include aerial photography and other documents that cover a large geographic
area and provide a complete record of changed mapping features. One example of periodic maintenance is updating buildings
and roads from new aerial photography in areas with substantial development. By comparing the base map to a new digital
orthophoto image, updates can be easily identified and changed. Periodic maintenance is often used for land base updates, or
when transactional maintenance is preferred but not practical because of the condition and accessibility of data sources.
Staffing
Staffing involves selecting appropriate personnel, deciding who's responsible for GIS maintenance, and training. Staffing
requirements depend completely on the amount of GIS maintenance required. This is sometimes difficult to determine when
planning a GIS; participants often do not know how much maintenance will be necessary until GIS development is under way.
However, GIS participants should still estimate their potential staffing needs while planning the maintenance program. These
estimates should be based on the key issues presented here: applications, accuracy levels, and the type of maintenance to be
used. Note that it's always better to overestimate rather than underestimate staffing needs.
The following decisions also should be made:
Maintenance Responsibility
Decision makers in multi-participant GIS programs usually assign the maintenance of specific, related classes of mapping
features to agencies that already maintain this data manually or to agencies whose GIS applications depend on this data.
For example, a planning department might maintain less accurate data such as zoning or land use data, while an engineering
department might maintain very accurate survey records for transportation data. However, even though these vary in levels of
accuracy, some communities have found it cost-effective to establish one department to oversee all maintenance activities, For
a maintenance program to be efficient and cost-effective, GIS decision makers should establish policies guaranteeing that
portions of the GIS being updated are restricted to read-only status until the changes are complete.
Maintenance Training
After selecting hardware and software for database maintenance, decision makers should determine who will be trained to
maintain the GIS. Will one department be formed that will be staffed by trained GIS personnel? If individual agencies are
responsible, decision makers should select those who are comfortable with and open-minded about GIS technology,
committed to using the technology, and willing to be trained.
Although it has become less rare to find people already familiar with GIS, a significant portion of the GIS maintenance budget
still must be dedicated to training. Even after training, it can take three to six months for a staff member to be technically
proficient with the equipment and software. For GIS maintenance to be cost-effective, decisions about staffing needs should
be made during GIS planning. If these decisions are postponed until maintenance is ready to begin, participants will likely be
faced with making expensive, laborious, and disruptive changes.
CONCLUSION
Decision makers should use the following guidelines when considering critical issues in GIS maintenance:
Implementing a GIS program is a long-term investment that requires participants to be dedicated and to commit their staffs and
budgets. But a GIS is only as useful as the data input. To protect their investment, participants must plan GIS maintenance
during database design, commit resources to implement a proper maintenance program, and establish procedures that will
cost-effectively keep the GIS up-to-date, complete, and accurate.
Unless participants have the ability and the budget necessary to properly maintain a GIS, they should think twice about
developing one at all. Implementing a GIS without maintaining it is a waste of time, money, and effort.
Creating and implementing a logical data maintenance plan is crucial so that a GIS is useful for years to come. Without proper,
planned data maintenance, the development of a GIS could become merely a one-time exercise in data conversion.
1. Manager, Business Development, Woolpert, 409 East Monument Avenue, Dayton, Ohio, 45402-1261.
home page