When SDS is processing records for sensitive species from iNaturalist, it is taking values, such as:
decimalLatitude
decimalLongitude
coordinateUncertaintyInMeters
-36.6381056263
149.5302313127
28505
and processing these to (actual output):
decimalLatitude
decimalLongitude
coordinateUncertaintyInMeters
-36.64
149.53
1000
In this case (catalog number: 107523407 - record link) the value of 1,000 for the processed record in incorrect, as the data is already provided with a larger value of 28,505.
Expected output is
decimalLatitude
decimalLongitude
coordinateUncertaintyInMeters
-36.64
149.53
28505
Background
The reason this is tripping up SDS, is the data is already generalised by iNaturalist but it uses an unorthodox method of generalising the coordinates. Instead of rounding the values, it uses an area of 28 km2 around the raw location and then radomly generates a new position inside that 28 km2 area.
SDS takes a conservative approach and does not just "trust" the already generalised flag on these records, as the data provider may use a generalisation method that is not as strict as the rules provided for SDS. So it performs its own generalisation on the coordinates, as seen above.
However, it should respect the provided coordinateUncertaintyInMeters and not use its own calculated value in this case, if the calculated values is less than the provided value.
Summary
When SDS is processing records for sensitive species from iNaturalist, it is taking values, such as:
and processing these to (actual output):
In this case (catalog number: 107523407 - record link) the value of
1,000for the processed record in incorrect, as the data is already provided with a larger value of28,505.Expected output is
Background
The reason this is tripping up SDS, is the data is already generalised by iNaturalist but it uses an unorthodox method of generalising the coordinates. Instead of rounding the values, it uses an area of 28 km2 around the raw location and then radomly generates a new position inside that 28 km2 area.
SDS takes a conservative approach and does not just "trust" the already generalised flag on these records, as the data provider may use a generalisation method that is not as strict as the rules provided for SDS. So it performs its own generalisation on the coordinates, as seen above.
However, it should respect the provided
coordinateUncertaintyInMetersand not use its own calculated value in this case, if the calculated values is less than the provided value.