Florida Museum of Natural History

University of Florida Herbarium (FLAS)

Map Image
Best Practices for Collecting Geographic Data in the Field
   by Joshua Robinson and Kent D. Perkins

(Beta Version 2, August 27, 2008)

Introduction

Detailed collecting locations are used by researchers for general mapping of collections; for detailed mapping, as in GIS computer applications; and to physically locate the plant(s) in order to obtain further research material. To facilitate research, locations should include country, state or province, county or municipality and a description of the location in reference to roads, road junctions, mile markers, geographic/geologic features and distances from cities and/or towns. Latitude and longitude, section, township and range, and elevation are also important. A location taken with a Global Positioning System (GPS) unit is now considered standard practice for research collecting.

It is important to have a standard way to collect geographic information in the field so that users of collections have the best possible information to base their research on. Collecting spatial data goes beyond coordinates, and includes issues such as extent and uncertainty. Physical locality descriptions are also important and provide an alternative for error checking. All of the information outlined in this guide will help others correctly georeference and use your collections in further research. Spatial data can be incorrectly manipulated if true standards are not followed. For example, in 1492 Christopher Columbus presents his argument to travel westward to Asia to the Spanish monarchs, Ferdinand and Isabella. Columbus knowingly uses Marinus’ calculation that the earth is only 25,255 km in circumference, despite experts believing Eratothenes’ calculation of 40,000 km. As the Spanish monarchs were not spatial experts, they accept Columbus’ proposal and send him westward to “Asia.” It is important that when collecting data, information is carefully recorded so that such deceptions cannot happen. The following manual will describe practices for collecting coordinates, location and mapping extent, and determining uncertainty.

Coordinates (Projection, Datum, Error)

In today’s world most spatial data is incorporated into GIS programs after being collected by GPS receivers. GPS use the term projection to refer to the coordinate system settings. It is suggested that the projection be set to lat/long (latitude / longitude)and that the values be recorded in decimal degrees. While there are many projection systems which points can be collected in, decimal degrees is the most precise for the curved nature of the earth. Record all decimal degrees provided by the unit. Decimal degrees are more readily converted into UTMs (Universal Transverse Mercator coordinate system) or other systems rather than conversions to decimal degrees.

A geodetic datum (datum) is a reference system by which measurements and coordinates are made. The earth is not flat, therefore many different methods for overlaying a grid and measuring system (datum) on its surface are devised for coordinates. Therefore, it is critical to know the datum for the latitude and longitude coordinates being collected. GPS systems should be set to the World Geodetic System of 1984 (WGS84) datum, unless the field location has its own best fit datum (e.g., Australia). Sometimes coordinates are calculated off of digital or paper maps. These values are also acceptable, especially from USGS topoquads or DOQQs, but the coordinate projection system and scale of the map should be reported. The datum is an essential data element that must reported as there are errors when converting coordinates between systems.

GPS receivers also add error to measurements. If your receiver and field location are capable of utilizing WAAS (Wide Area Augmentation System) or Differential GPS (DGPS), you should use these options to increase your accuracy. GPS points should not be collected if less than four satellites are being used (many newer units may not report values with less than four). It is important that when recording metadata for a collection the exact name and model of GPS unit is recorded as well as the average error given on the receiver. These, and other measurements, should be given in the metric system. Measuring altitude should preferably be based on the use of a barometric altimeter or a topoquad as GPS receivers have low vertical accuracy.

Location and Mapping Extent

Locations can be defined as points or areas. It is important to make this distinction when collecting. While it is preferable that coordinates are collected for the exact location of each specimen, it is acceptable to collect coordinates for a general area. Points should describe exact locations and should give a measurement value (in kilometers or miles), a direction, and a feature from where measurements are taken. Directions should be clearly identified as by road or air since there is a large difference between 4 miles north by road from the Hippodrome and 4 miles by air from the Hippodrome. By road the distance may not be a straight line, whereas by air it should be. Feature categories are also important. The collector should use their judgment when determining features to measure from. Acceptable features should be ones that are temporally static, such as major road intersections, and locations of permanent buildings. Even major natural features may not be reliable as they can change often. Describing a feature as a city is unacceptable since a city is not a point. It is important to name the city collecting is taking place in, but measurements should not be made from the “center” or “edge” of a city that can change drastically through time. As an example, compare the “northern edge” of the city of Gainesville in 1900 and 2000. Over 100 years the city has expanded with little sense of pattern or temporal stability. Collections from the 1950s that denote measurements from the “eastern edge of Gainesville” mean almost nothing today unless we search through 1950s maps. To avoid your collections from falling into this category, measurements should be made from major, permanent features, and not features like “10th St. construction site” or “25 meters from edge of forest,” although the latter may be included as part of the habitat information.

While it is acceptable to collect a single pair of coordinates for multiple specimens, say the center of a plot, it is important to note this fact and describe the extent of the plot. The coordinates should describe the center point of the feature and a radius measurement should be given that describes the boundaries. This extent is important to later users of the collection as it limits the possible uncertainty of the location of the collection. If a center point is collected for a collection in the Amazon, one should state that all specimens come from a 1 km radius around that point. Without this information, the extent could conceivably be from Marajo Island to the Andes. Such a collection would have little use if one is studying specific Amazonian regions.

Uncertainty

A general spatial uncertainty for the whole collection should be reported in the metadata. When determining an uncertainty many issues must be considered. Considering the GPS coordinates’ errors, extent given for a collection, distance measurements (4.2 km vs. 4.25 km), and directional offsets (NE vs. NNE) among other factors a robust uncertainty measurement can be determined. The collector themselves should determine this uncertainty as it is difficult for someone to pick out the spatial errors in another’s collection. Once the data is reported by the collector anyone who uses it considers it 100% accurate, or at least as described in terms of errors. It is better to overestimate this error rather than underestimate it. While there are numerous mathematical suggestions on how uncertainty can be calculated, there is little agreement at this point in time. The best suggestion is for the collector to use their own experience and information from the field to determine an acceptable uncertainty. It is preferable to have an uncertainty value that can be added to each point, for example +/- 5 meters, and to have a general uncertainty for the whole collection. When all specimens are collected in a general area an uncertainty circle should be described for the extent of the collection. Sources of error that add to uncertainty are expected and the collector should be honest about them. It is far worse to omit any errors as this severely reduces the value of the overall collection.

References

The above information is adopted from the following resource and modified specifically for our purposes. For more information or in depth topics please consult the following source:
Chapman, Arthur D. and John Wieczorek, eds. 2006. Guide to best practices for georeferencing. Global Biodiversity Information Facility, Copenhagen, [Denmark]. c2006, The Regents of the University of California. vi, 80 p. URL: http://www.gbif.org/prog/digis/Georeferencing

Field Checklist

This is a general checklist to what you should do in the field, it is by no means exhaustive and for more information the best practices guide should be consulted.