Geo-referencing tabs

At the North American User Group meeting in Chicago in 2005 the EMu Natural History Special Interest Group (NH SIG) proposed the design of a common tab for storing geo-referencing information. Discussions with a number of users took place over the next year and results were published on the EMuUsers forum for review. After considering all feedback the group produced a proposal and the recommendations were implemented in the Sites module and Collection Events module (EMu 4.0 onwards).

Changes to the Sites and Collection Events modules included the following:

• A common display tab and associated search tabs were designed and built.

  All clients who store geo-referencing data in either the Sites or Collection Events modules have had their modules adapted to use the new tabs.

• Support for multiple opinions was added, allowing data from various sources to be recorded (e.g. a set of co-ordinates from a GPS device, another set from a map reading and another set derived automatically from the Precise Location via BioGeomancer).
• A set of common columns was added to both the Sites and Collection Events modules.

  All existing data was migrated from client specific fields into the new common fields. The introduction of a core set of fields allows for easier data interchange.

• The calculation of centroid values was updated to use the new columns.

  The centroid computation is now performed via an external service (fifo server) and used by the Windows client and the database server.

• The precision was modified to imply more accuracy for all computations where the form of a Latitude or Longitude is changed (e.g. Degrees / Minutes / Seconds to Decimal).

The changes have led to the merging of many disparate versions of geo-referencing data within EMu client implementations. The new tabs should meet the needs of clients and allow for the creation of features based on a common framework.

EMu 3.1 saw changes made to the conversion of latitude and longitude values between decimal (Dec) and Degrees / Minutes / Seconds (DMS) formats. The rule applied, as used throughout EMu, was that a calculated value should never imply more precision than the original value. While this rule is useful in general it does cause problems with geo-referencing data. Using the decimal representation the precision of a latitude / longitude value is 1/10, however for DMS the precision is 1/60, so when converting from DMS to decimal noticeable precision is lost. If a decimal precision of 1/100 is used, a slight gain in precision over DMS occurs.

The EMu NH SIG recommended a change in how precision should be handled when converting between latitude / longitude values in DMS and decimal formats. The recommendations were:

• If a DMS value is entered, the decimal value will be converted to two decimal places if seconds is missing, four decimal places if whole seconds is provided, five decimal places if one second decimal place is provided, six decimal places if two seconds decimal places are provided, and seven if three seconds decimal places are provided, etc.
• If a decimal value is entered, the precision of the DMS should be calculated in reverse of the above. Two decimal places round to minutes, three or four decimal places round to seconds, five decimal places round to tenths of a second, six decimal places round to hundredths of a second, seven decimal places round to thousandths of a second, etc.
• Centroid calculations should be given one degree of precision greater than the least precise latitude/longitude value in the group. Although this may be implying an increase in degree of accuracy, it is only one degree and will help lessen the skewing of the centroid values. The centroid calculator does not imply extra precision where a single latitude / longitude pair is specified.

Each of the above recommendations has now been implemented.