yes

Hi Dr. Zheng,

In my opinion, you can use species' latitudinal midpoints to distinguish them on broad spatial scale, e.g. polar vs temperate vs tropical. But it doesn't say anything about the width of the said distributions. While 2 species can have similar midpoints, one may present a narrow range whereas the other is ubiquist. So it is as important to determine the latitudinal width of the species ranges.

Technically speaking, just add the lowest latitude to the highest one and divide by 2.

You should also use percentile distribution to avoid using the most extreme records. And assessing the shape of the latitude data distribution might also be necessary, depending on how many points your species have.

Instead of midpoint, you should also try the median, but some studies have identified the midpoint as more robust (i.e. less sensitive to sampling effort and distribution data).

Best,

thanks all, it is very helpful indeed.

It is clearer to me how to use the midpoint data. I am going to assess the distribution of latitude data and the width, but still no sure how to use the percentile distribution, is it mean that I should remove the 5% of the lowest and highest value for a species'  location data?  is there any paper to referred to? 

once a while, I was thinking may centric coordinates for all the GBIF data is better, however, it is not so easy to do it right as some of the species distributed as clusters in several regions, such as East and West of USA or CHINA, which is mean the centric of them is in the middle region, certainly this is not right, as there are quite different climate in three regions. maybe only mid-latitude is meaningful in such scenario.

hoping for further comments.

best wishes

Jingming

Hi Dr Zheng,

Unfortunately you will rarely achieve a reasonable answer when using an average, median or other summarising statistic unless you have a naturally occurring normal distribution that has been either truly randomly sampled or geographically systematically sampled. This is rare. Without assessment of the sampling density you cannot know whether the points represent the species distribution. You can get weird results e.g. a terrestrial species around the edge of a bay would have a centroid that is over water and an aquatic species that is sampled around an island would be over land. Unless this is a rapid and rudimentary study I would encourage you to use all available points (remove duplicates where possible). If you must use summarised data, then have a good look at the distribution and possibly hand select (move the centroid or average) the point manually to better represent the environmental conditions. If possible separate the data sources and see whether they give the same broad distributions. If one is different for a particular location, this may be due to a misidentification, or due to a species name change etc.

Best wishes

Chris Allen

good points and thanks, Chirs.

I do check the source of the records and found some errors in GBIF data. another options I used is to compare GBIF with local data and species distribution maps, eliminate the unreasonable data manually. so far, it seems that the midpoint of latitude is relatively reliable, but not longitude, so I only use the mid-altitude data for a large scale analysis along with other variables. in this case, I would not use the calculated coordinates to retrieve climate information.

There are certainly some problem in GBIF information, but no data is perfect, isn't it? we just need to make a good protocol agreed by communities. are there such thing? 

Jingming

https://rdrr.io/github/macroecology/letsR/man/lets.midpoint.html