Uncertain Environmental Variables in GIS
- PDF / 223,949 Bytes
- 6 Pages / 547.087 x 737.008 pts Page_size
- 42 Downloads / 177 Views
UMN MapServer
UMN MapServer University of Minnesota (UMN) Map Server
Uncertain Boundaries Objects with Broad Boundaries
how much is known about them. Uncertainty can be conveniently represented using probability theory, which provides a sound theoretical basis for assessing and propagating uncertainty. When probability distributions cannot be estimated, less detailed approximations may nevertheless be possible. These include intervals and scenarios, where possible outcomes are listed without their associated probabilities. Historical Background
Uncertain Environmental Variables in GIS H EUVELINK 1, JAMES
B ROWN 2
G ERARD B.M. D. Environmental Sciences Group, Wageningen University, Wageningen, The Netherlands 2 National Weather Service, National Oceanic and Atmospheric Administration, Silver Spring, MD, USA 1
Synonyms Spatial data quality; Attribute and positional error in GIS; Spatial accuracy assessment; Accuracy; Error; Probability theory; Object-oriented; Taylor series; Monte carlo simulation Definition Environmental variables are inherently uncertain. For example, instruments cannot measure with perfect accuracy, samples are not exhaustive, variables change over time (in partially unpredictable ways), and abstractions and simplifications of the real world are necessary when resources are limited. While these imperfections are frequently ignored in GIS analyses, the importance of developing ‘uncertainty aware’ GIS has received increasing attention in recent years. Assessing and communicating uncertainty is important for establishing the value of data as an input to decision-making, for judging the credibility of decisions that are informed by data and for directing resources towards improving data quality. In this context, uncertainties in data propagate through GIS analyses and adversely affect decision making. Error may be defined as the difference between reality and a representation of reality. In practice, errors are not exactly known. At best, users have some idea about the distribution of values that the error is likely to take. Perhaps it is known that the chances are equal that the error is positive or negative, or it may safely be assumed that in only one out of ten cases the absolute error is greater than a given number. For variables measured on a categorical scale, certain outcomes may be assumed more likely than others. Thus, uncertainty stems from an acknowledgement that errors may exist, and its magnitude is then proportional to
Interest in the problem of error and uncertainty in GIS dates back to the early years of GIS. For example, Peter Burrough devoted a chapter to the subject in his 1986 book on GIS. Burrough presented an exhaustive list of possible sources of error in spatial data and analysis. He also demonstrated the use of simple arithmetical rules from error analysis theory to propagate errors through GIS operations. Shortly after, the National Center for Geographic Information and Analysis in the USA defined uncertainty in spatial databases as one of its key research themes
