Marine Minerals Research In The United States: Recent Results From The Marine Minerals Technology Center

The Marine Minerals Technology Center (MMTC) was authorized by an act of the United States Congress in 1996. The act established several marine minerals research centers in the U.S., including one at the University of Alaska Fairbanks. The Alaska MMTC is funded through the Minerals Management Service of the Department of the Interior and concentrates its research in the Arctic seas region. Recent research projects at the MMTC include, among others, application of a geographic information system (GIS) to the gold placer deposits offshore Nome, Alaska, development of a marine gold placer reserve definition model using a neural network, and development of an expert system model for submarine tailings disposal (STD) planning and decision-making. Although geostatistics remains the most prevalent method used today for ore grade estimation, researchers have made numerous improvements over the years for accurately predicting grades of ore using other estimation techniques. One such technique, the neural network, has recently emerged as an alternative to geostatistics for grade estimation purposes. A neural network is a non-linear, model-free estimator that is well-suited for noisy and/or sparse data environments. Neural networks perform best with non-linear spatial data, contrary to the stationary assumption of ordinary Kriging techniques. In the marine placer ore reserve definition model, the neural network technique was used to estimate placer gold reserves offshore Nome. This study demonstrated the utility of neural network modeling over other traditional ore reserve estimation methods, especially for high cut-off grades. Some of the relevant issues of neural network modeling were also addressed in the study. In testing, the neural network produced results that were close to those from geostatistical techniques including Kriging, polygonal, and inverse distance methods.