Hello Somaye,

I'm not quite sure that I understand your question but I think you are searching for a way to:

1. combine 2 sets of data with different spatial resolutions in the same area and

2. interpolate the primary variable using a secondary variable (probably with a better spatial resolution).

You have some choices:

1. cokriging > you can perform full block cokriging and/or Markov model cokriging. I suggest you to use the first one because of some reasons. cokriging provide best estimates (minimum error) and estimation error at every block but it can't provide a way to assess the uncertainty. you can perform "Indicator cokriging" to have this.

2. Co-simulation (Co-SGS or direct co-simulation (Co-DSS)) gives multiple realizations which can be used to have a vision on the variable fluctuation. you can use simulations to model the uncertainty.

3. Regression kriging > it is easier rather than previous methods but has certain limitations.

for more details, please see the links.

I suggest you to tag your question with "geostatistical analysis" to have a wider range of answers from more experts.

http://www.sciencedirect.com/science/article/pii/S1574954107000386

http://www.sciencedirect.com/science/article/pii/S0098300404001414

What kind of plot data do you have? Is it a field-based inventory? You approach may depend on your data.

Before spatial interpolation, you may have to estimate tree biomass - how to do that depends, one more time, on you plot data.

To make a accurate kriging Interpolation , I think you need more sample than 7 even your study area about 1 sq km. another thing about distribution of sample in the study area .  

Hi, nice to see that more people are starting to use UAVs as sampling tools as done in our recent paper (http://www.sciencedirect.com/science/article/pii/S0034425717301220). I also worked on a similar issue that you are addressing, i.e. using three levels of information (field data, UAV, and satellite imagery) and I hope it will be available in the near future.

For some more info on geostatistacal methods when you want to include more levels of information you can see this paper from some colleagues (https://arxiv.org/abs/1705.03534) or the work done By Saarela et al on hierarchical model-based approaches https://link.springer.com/article/10.1007/s13595-016-0590-1

Cheers, 

Stefano

http://www.sciencedirect.com/science/article/pii/S0034425717301220