GEOQUALv2: Advancing Georeferencing Quality Analysis

With the launch of the new version of the CAPFITOGEN3 (30/07/2024) toolbox scripts, we have aimed to strongly promote the GEOQUAL tool, which, since its release in 2013, had undergone almost no changes. We can say that GEOQUAL has been doubly renewed. First, let’s explain the changes the original version of GEOQUAL has undergone. Among the parameters measured by GEOQUAL concerning the quality of georeferencing of (plant) occurrences or collecting sites of accessions, this tool is based on three general criteria and several sub-criteria. The criteria are:

• Suitqual: Evaluates the quality of georeferencing regarding how suitable the site is for the growth of a plant (cultivated or wild, depending on what the user indicates in the passport data in the SAMPSTAT field). It was originally based on the Global Land Cover map (Globcover), which at the time was available with global coverage, cell size of 1×1 km, and with the oldest data (land use for the year 2000). The most comprehensive, valid, and oldest map possible was preferred because most of the germplasm/plant collections/occurrences were before the year 2000 at that time (2013).
• Localqual: Evaluates the correspondence between the fields describing the collectimng site/occurrence (ORIGCTY and ADMs fields in passport data format) and the locality data from the Global Administrative Areas Database GADM version 2.8 (available 13 years ago), extracted for the coordinates indicated in the passport data (DECLONGITUDE and DECLATITUDE fields).
• Coorqual: Quality of the coordinates indicated in the passport data based on some sub-criteria extracted from the information provided by the user, explained in the CAPFITOGEN3 user manual (see manual).

The information from these three criteria is summarized in a final quality indicator called TOTALQUAL, which could be used to filter the records with the highest quality in their georeferencing. The original GEOQUAL tool had the problem that even if the passport data contained recent occurrences or collection sites (according to the COLLDATE descriptor), the Suitqual criterion was obtained from 1×1 km cell maps from the year 2000, and the Localqual criterion from administrative area maps from 2010 or earlier, which had often been modified later. Based on this and given the greater computational capacity of current computers, GEOQUAL was one of the tools that needed to be updated.

The GEOQUAL tool update includes:

• Elimination of the “precision” parameter since all evaluations from 2024 onwards are made on high-resolution coastline maps.
• The global land cover map (Globcover) for determining Suitqual remains for the year 2000, but its resolution increases from 1×1 km to 300×300 m.

However, the new GEOQUALv2 tool performs a more complete and accurate analysis than the updated version of GEOQUAL, including:

• The Suitqual criterion is calculated for each collection/occurrence site according to the year the germplasm or plant was collected/observed, indicated in the passport data in the COLLDATE column as follows: from 1996 and earlier, a Globcover map from 1992 is used; from 1997 to 2008, a Globcover from 2000; from 2009 to 2019, a Globcover from 2015; and from 2016 onwards, a Globcover from 2022. All these land use maps have a resolution of 300×300 m. This tool, like the original version, also applies a suitability table (see) for each original land cover code, differing if it is a cultivated or wild species.
• The Localqual criterion is also determined for each collection/occurrence site according to the year the germplasm or plant was collected/observed, indicated in the passport data in the COLLDATE column as follows: from 2015 and earlier, using the GADM 2.8 version (see) and for more recent records, the current GADM 4.1 version.

With these improvements in version 2 of GEOQUAL, we hope to offer CAPFITOGEN3 users much more valid results in line with the most current geoinformation. The greater capacity to offer more precise and current results comes at a cost: the CAPFITOGEN3 toolbox installer increases from 1.9 GB to approximately 6.1 GB.

This installation file size and the greater computational capacity required to run this new GEOQUALv2 would have been a significant limitation 13 years ago when the original GEOQUAL tool was launched. However, by 2024, current internet speeds and accesibility and computer capabilities do not represent a real limit for downloading and using this new version of the tool.