These GIS files (UTM30N WGS84 Shapefiles and GeoTIFFs) of gridded fine-scale harbour seal (Phoca vitulina) at-sea (i.e. marine) usage estimates for a region around Orkney and the North Coast of Scotland were produced for Marine Scotland (MSQ-0174, Contract Research Fund - Ref CR/2014/11) by the Sea Mammal Research Unit (SMRU). An R-based modelling process was used to combine harbour seal telemetry data from 60 adult (>1-year-old) harbour seals which spent time within the study area, terrestrial count data, and environmental covariates to produce an estimated at-sea usage and associated 95% confidence intervals. These predictions can be interpreted as the average number of harbour seals in each 600 m by 600 m grid cell at any point in time. The kernel smoothing used to change the point locations to a usage surface can create at-sea usage inland, so to accurately sum usage in an area, ensure that all local onshore and offshore grid cells are summed. The telemetry data for 20 of the seals were obtained from SMRU Instrumentation Satellite Relay Data Loggers (SRDL) deployed over the years 2003-2005. Those SRDLs used the Argos satellite system for data transmission and to provide the location estimates. The telemetry data for the other 40 individuals came from SMRU Instrumentation Fastloc© GPS GSM “phone” tags deployed between 2011 and 2015. Locations were processed through a set of data-cleansing protocols to remove null and missing values and any duplicated records. The SRDL positional error was corrected using a Kalman filter and data were used to estimate positions at two-hourly intervals. The majority of GPS locations have an expected error of = 55 m; occasional outliers were excluded using thresholds of residual error and number of satellites, and then straight-line interpolated to regularise to the same two-hourly intervals as the SRDL data. Three animals had very few locations within the study area and three others had no haulout records. Those six animals were therefore excluded from the analysis, reducing the total number of animals used to 54. The terrestrial count data came from aerial surveys of harbour seals carried out by SMRU during the annual August moult. During such surveys, all seals along a specified coastline are counted and their coordinates recorded to an accuracy of approximately 50 m. Surveys take place within two hours of low tide when low tide is between 12:00 and 18:00 hours. Survey effort was variable between locations. Surveyed coastline was gridded to 600 m x 600 m and the most recent available count (ranging from 2008 and 2015) was used in each grid cell. Grid cells that were surveyed but in which no animals were located were given a value of 0. The environmental covariates used in the modelling were geodesic distance to haulout (calculated using GSHHS shorelines); gridded seabed depth (downloaded from the EMODnet bathymetry portal); proportion of sediment type (British Geological Survey); and annual mean tidal power and peak flow for mean spring and neap tides (PFOW Hydrodynamic model). Full details on the methodology used to produce the usage estimates from these datasets are given in the associated Scottish Marine and Freshwater Science Report Vol 7 No 27: "Fine-scale harbour seal at-sea usage mapping around Orkney and the North coast of Scotland" by Esther L. Jones, Sophie Smout, Clint Blight, Carol Sparling and Bernie McConnell.