Lec 12: Map Accuracy

Geographic Flight Simulation — A free form discussion

Virtual Georeconnaissance

Quote:

The Readings for Today’s Lecture were:

Ch. 20: Map Accuracy

Comments or Questions?

(General, Lab, Assignment, Readings)

Today’s Lecture:

Map Accuracy

The Course in Review

Map Accuracy

* TOC

Maps are a representation of reality

map accuracy refers to how true a map is to reality

map accuracy is a complex issue

TRUTH OR EFFECTS

“Our big concern

must be to compensate for the distorting or untruthful aspects of abstraction. This is true of all forms of representation (including mapping)”

Accuracy or Precision?

Accuracy means fidelity to the truth

Precision means “how fine is the measurement”

therefore high precision does not guarantee high accuracy

Factors involved in precision

measurement resolution

example: can a compass measure up to single degrees or tenth of a degree?

repeatability or agreement of multiple measurements

eg. multiple measurements taken with a GPS at the same location

the rigor and sophistication of the measurement process

example: modern telescope technology (post Hubble), including digital processing to compensate for atmospheric effects

Types of Accuracy

Positional Errors

horizontal or vertical

Question: Compared to What?

Answer: to US Map Accuracy Standards

refer to Box 20.1

Map accuracy is defined statistically

based on a small sample of locations

dangerous to deduct a specific instance (a particular location) based on the statistic(s)

positional map accuracy increases as map scale increases

Conceptual Effects

Question: what is meant by this?

Answer: conceptual (abstract) environmental features

e.g. “forest patch” “corridor”

see Box 20.2

Standards cannot apply to these types of features

which are quite different from base map themes such as topography, hydrology, roads, boundaries, and other

Generalization Effects

The mapping process demands cartographic generalization

Why?

This data reduction process results in loss of information and positional distortion

this positional displacement should be regarded as an effect rather than as error

within the same map a country road may be less generalized than a more intricate feature, such as a major highway

* Fig 20.1

Factual Errors

Question: What are these? Examples?

Note

that where remote sensing image processing is concerned

we are dealing with classification accuracy

errors of omission and comission

the concept of “mixels” (mixed pixels)

and sub-cell image processing

Timeliness

maps become “senile” over time...

like milk, maps are best ‘consumed’ when fresh...

note that different features in maps (esp. digital) may come from variably-aged data

COMMUNICATING MAP ACCURACY

It is important

that the map-maker communicate map accuracy to the map-user

Active Mapping Strategies

The map-maker tailors the map with map accuracy in mind

Adjusting Map Scale

if some of the data is inaccurate relative to the map scale, then the map scale should be decreased

Adjust Generalization

“As a rule, all variables on a map should be about equally generalized.”

In practice however, base data is often less generalized than less accurate thematic overlays

therefore avoid assessing the thematic map’s accuracy based on the base data

User-Oriented Strategies

When map-makers don’t actively incorporate accuracy considerations in their map design

they at least often take a more passive route and warn users

three types of warnings

Symbols and Notations

special symbols (explained in the map legend)

using a symbol fuzziness factor

Legend Disclaimer

placing a disclaimer in the legend

Example: the disclaimer that appears on the screen of the Garmin GPS after you turn it on:

“Land data is provided only as a general reference to your surroundings”

In addition

Maps often report the date of map production

and the source of the data

varying reputation of mapping agencies

commercial more suspect

Reliability Diagram

is sometimes included

* Fig. 20.2 provides examples of map accuracy warnings

DIGITAL RECORDS

digital maps come from secondary and primary data collection

Geographic Database

Primary data collection by the map-maker

especially in the age of GPS

has the potential for higher data fidelity

Primary data can be termed geographic data

that is compiled into geographic databases

Cartographic Database

The vast majority of digital spatial data and databases

Building digital spatial databases is hard

“Expending vast amounts of time, energy, and money doesn’t guarantee that a reliable cartographic database will be created.”

The map accuracy problem is compounded

when in addition to the existing problems with the data

more error is introduce in the process of converting this data into digital form

* For example, see Fig. 20.3

One classic problem

is the edge-matching problem when creating a ‘seamless digital database’

* Fig. 20.4

Data Lineage

A geneaology for maps

showing their ‘pedigree’

a bit like family trees for people...

Meta-data

Data about the data

standards exist in some countries and mapping agencies

on how data lineage should be documented

There are software modules and packages to deal with cartographic data lineage

eg. David Lanter’s Ph.D. and subsequent work on the Geolineus system

MAP COMPOSITING

Map Overlay

overlay logic and arithmetic is used on a number of layers to produce a result

cartographic accuracy suffers from additive layer error effects

therefore results must be treated with special caution

* Fig. 20.5

Cartographic Modeling

The problem gets worse when more sophisticated modeling procedures are used, due to:

selective choice of data layers that are included in the model

weights assigned to variables

the mathematical model itself

Example: the Unversal Soil Loss Equation (USLE)

E = 2.24 * R * K * L* S* C * P, where

E - erosion (average soil loss), R - rainfall intensity factor, K - soil erodibility factor, L - length-of-slope, S - steepness of slope C - cropping management factor, P - prevention practice factor

[http://pasture.ecn.purdue.edu/~aggrass/models/agnps/usle.html]

Sensitivity Analysis

aka Stability Analysis

parameters are slightly changed

and the results compared

if the results vary wildly, the model and maps are deemed highly sensitive (unstable)

and should be given less credibility

LIABILITY ISSUES

facing map-maker and map-user

today often the map-user is also the map-maker

so both liability issues may pertain to the same individual or organization

Map-Maker Responsibility

must make “every reasonable effort” to ensure map or cartographic database quality

must warn user of potential problems and hazards in using the product

Map-User Responsibility

less well defined

if you are a vendor and market map-use products

you should be careful and responsible

“Speaking ethically rather than legally,

you probably have no business using maps if low skill on your part could harm others or the environment.”

You especially have an obligation to get beyond “map-as-territory” blinders.

Question: Huh? What does this mean?...

CONCLUSIONS

Map user should be aware of map accuracy issues and ask:

“Can I live with the accuracy of this map in this situation?”, or

“What special care should I take if I use this map for this purpose?”

Break

The Course in Review

End of Course : )