Measaured samples

[anquan xia]

Dear WOSIS Administor,

    Nice day~

    Considering the particularity of the study area, I want to use measured samples/profiles of SOC density(0-30 cm) from WOSIS and combined our field measured data to model the SOC density around the Himalaya area. but I cant find the measured SOC density information could be used directly. Can you share some experences to me about how to get these samples?

     I have 2 thoughts. Can you help me check the feasibility and precision?

     ①I use the "wosis_latest_profiles" point layer and according the samples position to extract the SOC density from soilgrid250m image, and use this extract data as true or measured data.

     ② In the soilgrid250m web,  I use the uncertainty layer (e.g. uncertainly<5 or 10 )to filter the imge, then use the rest data as true or measured data.

     Can you give me some suggestion or advice? 

     Thank you,

     Best regards,

     Anquan

[Niels Batjes]

WoSIS does not contain data on SOC density. You will have to calculate this from the data on SOC, BD and coarse fragments and layer depth.

You should not use the point data location to extract SOC density from the SoilGrids250m layers! 

Op woensdag 4 november 2020 om 09:36:53 UTC+1 schreef xiaan...@mails.ucas.ac.cn:

[anquan xia]

Can you give me some reason? why i cannt use the point data location to extract SOC density from the SoilGrids250m layers ?  In my opinion, the random forest is base on the point data, so i think the SOC density or SOC storage is more accurate in the measured point location than other location.

Thank you again,

Anquan