The Fennoscandian (or Baltic) Shield represents the largest outcropping domain of Precambrian bedrock in Europe, covering more than a million km2 throughout Norway, Sweden, Finland, and northwestern Russia. This book focuses on Finland, which occupies the central part of the shield and which, since the advent of modern geology in the 19th century, has been instrumental in a number of fundamental insights and advances in understanding Earth processes.

Precambrian mineral deposits are now becoming more significant in the mineral-raw materials balance for many countries in the world, since the vast majority of ores of iron, titanium, vanadium, gold and uranium, non-metals such as muscovite and phlogopite, and considerable amounts of asbestos and barite are currently being worked from Precambrian deposits. Recently, new types of gold, rare-metal, tungsten, tin, beryllium, rare-earth, radioactive elements, platinoids and manganese ore bodies have been discovered, as well as deposits of unusual geochemical associations such as Ni-U, U-Cu-Au and Pt-Au.

This book got its start about seven years ago when Robert Bakker-the progressive dinosaurologist I studied with for many years-suggested that I draw up illustrations of some predatory dinosaur skeletons. About the same time, I started looking into the problem of bird origins, which I believed was tied to dinosaurs. One thing led to another and after a while I realized I had enough material available to put together a book on the dinosaurian meat eaters.

Modelling and prediction of spatially distributed data such as the secondary cassiterite mineral distributions are often affected by spatial autocorrelation (SAC); a phenomenon that violates attributes data independence in space, which leads to type1 errors in classical statistics and overfitting or underfitting in machine learning (ML) classification respectively. The concept of overfitting and underfitting of spatially distributed datasets in an ML classification has not been properly addressed by the traditional random holdout technique of model validation, and this is a challenge to the assessment of predictive spatial model performance in spatially distributed datasets.