This article is a continuation of the two previous ones: Let's hit the bispectrum off-road, or how to find gold in Siberia , in which we examined the geological model of a gold deposit in the Novosibirsk region and We are looking for ore gold on the island of Sumbawa, Indonesia , in which we built a geologically driven machine learning model to search for gold or other ore minerals throughout the Pacific ore belt, using open data for geological modeling on the Google Earth Engine (GEE) platform.
Having convinced of the similarity of the geological models of Siberia and Indonesia, we apply the classifier for ore gold of West Sumbawa, Indonesia for Siberia. In this way, we will receive a geologically correct detailed forecast of the gold content for Siberia, without using any geological data for this region at all.
Comparison of models in the territory of Indonesia and Siberia
We already have a classifier for Indonesia trained on detailed open data, which we would like to use in Siberia. Let us compare the detailed density models without taking into account the influence of the relief, constructing a model for Western Siberia with the same parameters as for Western Sumbawa and choosing equal areas. On the left model, dots on the surface show wells with high (Au) and low samples of gold (Au), and on the right contours, known regional gold (Au) and molybdenum (Mo) deposits are displayed:
, . , . , .
, :
— , . , .
(, ):
, . 3D :
, , . , , . , N-45-XIII. , - 1:500 000 - . , ( , ) — , , ( , ) , .
Google Earth Engine (GEE) , , , , , .
, , , , .