Most of the literature I read primarily focuses on 10,000 data points.

Dear Iresha,

This is not a "single answer" type of question, given that SDMs and ENMs tend to be tailored to specific study areas or research objectives. That said, if you pretend to "deviate" from the default 10,000 I'd recommend a read of some papers that discuss background data:

• Phillips and Dudík 2008 - Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation
• Barbet-Massin et al 2012 - Selecting pseudo-absences for species distribution models: how, where and how many?
• Whitford, Shipley, and McGuire 2024 - The influence of the number and distribution of background points in presence-background species distribution models

Also, I'd recommend a good read of Zurrell et al 2020 - A standard protocol for reporting species distribution models - given that it's one of the best protocols around when dealing with SDMs. Other considerations you might take in your research relates to the species distribution, how many occurrence points you have, data clustering and so on.

I hope that my answer may have shed some light in your question but feel free to contact me if you want to discuss it further.

Best wishes,

Gustavo R. Brito

Thank you Marcelo and Gustavo, your solutions are really helpful.

Gustavo Reis de Brito the study is about multiple species. Previously I used 10,000 but considering the size of the country, it was suggested to minimize! question is how. I've tried with the ratio to the occurrence as usually mentioned in literature such as 1:10 and 1:5 with the 10km buffer to reduce geographic bias, but many results were weird. Then used 1000, and seems results bit promising than before. Seems Barbet-Massin et al 2012 and Whitford, Shipley, and McGuire 2024 publications indeed. :)