The improvements of the Costa Rican ground-based gravity dataset as a result of a comprehensive attribute and spatial assessment of the historical databases

Abstract

Improving the accuracy of gravity models and their derived products such as Geoid models requires not only accurate airborne, shipborne, and satellite gravity measurements but also a good distribution of ground-based gravity measurements. Since terrestrial gravity surveys are limited by access and may entail a great cost, scientists rely on historical data generated by past surveys. Such datasets are heterogeneous and often lack meta-data. This work achieved a homogenized gravity dataset for Costa Rica and characterizes the spatial distribution of existing, good-quality data in order to consider the focus of terrestrial gravity surveys. First, the gravity value quality control removed duplicates and identified outliers through a process of leave one out cross-validation. Then, we compared terrestrial gravity disturbance values with data from Combined Global Geopotential Models such as EGM2008, GECO, EIGEN-6C4, and the SGG-UGM-1 for quality assessment and consistency of the data. Furthermore, we evaluated a 5′ grid for Costa Rica considering the best spatial resolution expected for the current Combined Global Geopotential Models. Taking into account the spatial distribution of the validated historical dataset, we surveyed new gravity points to significantly improve the coverage, reducing the 5′ × 5′ cells lacking data from 30% to 15%. This resulted in a complete dataset on a 5′ grid except for areas that are exceedingly difficult to access such as national parks, indigenous areas, and restricted private properties.