1

Topic: Point-Buffer - crazy big results - is it me?

Hi,

I'm new to exploiting the NBN webservices so you will have to excuse me if I am doing something dumb.

I want to characterise an area by doing a point buffer query to get the number of species ever recorded there.

But the results I get back are very high!  13,778 records and 3,574 species for a 100m radius circle in Edinburgh - which seems a little excessive. I wasn't expecting to get anything.

I paste below the query URI and also the non URL encoded polygon.

I'd be really grateful for any advice. Am I doing something stupid with my query construction or is the service resolution only to recorder district or something like that.

Many thanks,

Roger

p.s. the forum won't let me post links so I have removed the h t t p s

://data.nbn.org.uk/api/taxonObservations/species?absence=false&polygon=POLYGON%28%28-3.198712+55.966303%2C+-3.198434+55.966289%2C+-3.198163+55.966249%2C+-3.19791+55.966183%2C+-3.197681+55.966093%2C+-3.197483+55.965982%2C+-3.197322+55.965854%2C+-3.197204+55.965712%2C+-3.197132+55.965561%2C+-3.197107+55.965405%2C+-3.197132+55.965249%2C+-3.197204+55.965098%2C+-3.197322+55.964956%2C+-3.197483+55.964827%2C+-3.197681+55.964717%2C+-3.19791+55.964627%2C+-3.198163+55.964561%2C+-3.198434+55.96452%2C+-3.198712+55.964506%2C+-3.198991+55.96452%2C+-3.199261+55.964561%2C+-3.199515+55.964627%2C+-3.199744+55.964717%2C+-3.199942+55.964827%2C+-3.200102+55.964956%2C+-3.200221+55.965098%2C+-3.200293+55.965249%2C+-3.200317+55.965405%2C+-3.200293+55.965561%2C+-3.200221+55.965712%2C+-3.200102+55.965854%2C+-3.199942+55.965982%2C+-3.199744+55.966093%2C+-3.199515+55.966183%2C+-3.199261+55.966249%2C+-3.198991+55.966289%2C+-3.198712+55.966303%29%29

POLYGON((-3.198712 55.966303, -3.198434 55.966289, -3.198163 55.966249, -3.19791 55.966183, -3.197681 55.966093, -3.197483 55.965982, -3.197322 55.965854, -3.197204 55.965712, -3.197132 55.965561, -3.197107 55.965405, -3.197132 55.965249, -3.197204 55.965098, -3.197322 55.964956, -3.197483 55.964827, -3.197681 55.964717, -3.19791 55.964627, -3.198163 55.964561, -3.198434 55.96452, -3.198712 55.964506, -3.198991 55.96452, -3.199261 55.964561, -3.199515 55.964627, -3.199744 55.964717, -3.199942 55.964827, -3.200102 55.964956, -3.200221 55.965098, -3.200293 55.965249, -3.200317 55.965405, -3.200293 55.965561, -3.200221 55.965712, -3.200102 55.965854, -3.199942 55.965982, -3.199744 55.966093, -3.199515 55.966183, -3.199261 55.966249, -3.198991 55.966289, -3.198712 55.966303))

2

Re: Point-Buffer - crazy big results - is it me?

Hi Roger,

Your results will include all records that overlap the specified polygon. If there are datasets with 10km public viewing for that area, or 10km records within datasets with finer-resolution public viewing, you may be getting records from up to 14km away included in the summary. You may want to check if this is the case for you.

The web-services for Gateway 5 have no solution for this (unlike Gateway 4 which allowed a maximum resolution for included records to be specified). To get around this problem I have had to use api/taxonObservations to download all the records and then sift-out all the coarse 10km records before building my own species summary from what's left.

As I also want taxon group and common names in the species summary I still have to use api/taxonObservations/species to get these (which aren't provided by api/taxonObservations), then stitch the two sets of information together!

Hope that helps?

Keith

3 (edited by charliebarnes 12-10-2015 09:00:02)

Re: Point-Buffer - crazy big results - is it me?

I may be wrong, but doesn't the taxonObservations API default to include overlapping grid squares? That means you'll be getting back records for NT2575 as well as NT25257539. That may explain the large number? Especially if there are a lot of records at 10km 'Edinburgh'.

Edit: thanks Keith!

Charlie Barnes
Information Officer
Greater Lincolnshire Nature Partnership

4

Re: Point-Buffer - crazy big results - is it me?

Thanks Charlie,

I guess if I can't do down to about 1km square it isn't worth doing for my project. Is there a simple way I could get a 1km square species list from a lon/lat?

Roger

5

Re: Point-Buffer - crazy big results - is it me?

can you use &spatialRelationship=within or something similar to overrule the default =overlap. Might end up with 0 results though

Gordon Barker
Biological Survey Data Manager
National Trust

6

Re: Point-Buffer - crazy big results - is it me?

Brilliant. Thanks Gordon. Just what I needed to know. If I include spatialRelationship=within and increase my radius to 500m then I get point "a few" records areas like Local Nature Reserves, that one would expect to get data for, and none for other areas. This is what I was after.

7

Re: Point-Buffer - crazy big results - is it me?

You may not get much useful data unless you know that the public viewing access is set to 100m for the datasets you are interested in and the records they contain have 100m resolution. If public viewing is set to 1km or coarser resolution, or the records have been recorded with 1km or coarser resolution (which is often the case for common species) you will not get these records until the whole of their squares fall completely within your polygon. You may therefore need to increase your buffer size still further, but then you will get more 100m records from outside your area of interest. Swings and roundabouts really.

In essence it isn't possible using gridrefs alone to get a good species list for a site. You need to look at location descriptions in the records too (if they are visible).

You can also ask for enhanced access if you want to use records at the resolution they were recorded, which may reduce the quantity of "distant" records that are fogging the picture.

Keith