Spatial modeling and uncertainty characterization of polymetallic nodules in the Clarion-Clipperton zone

Modeling the spatial variation of resources is necessary because it gives an estimate of what to expect during their exploration and exploitation. We focus on the spatial modeling of polymetallic nodules found in the deep sea regions of the Clarion-Clipperton zone in the Pacific. The data from this region available in the open domain is sparse, which warrants modeling techniques that can efficiently use the data to yield reasonable estimates. Additionally, it is necessary to quantify the confidence level of the model's predictions to aid a user in interpreting them. The confidence level can be captured in the form of an uncertainty map that accompanies the prediction map. We present an approach for modeling of copper percentage variation in nodules found across the Clarion-Clipperton zone. We also describe how the prediction uncertainty given the input features can be estimated.