This dataset encompasses data produced in the study ''Seasonal Arctic sea ice forecasting with probabilistic deep learning'', published in Nature Communications. The study introduces a new Arctic sea ice forecasting AI system, IceNet, which predicts monthly-averaged sea ice probability (SIP; probability of sea ice concentration > 15%) up to 6 months ahead at 25 km resolution. The study demonstrated IceNet''s superior seasonal forecasting skill over a state-of-the-art physics-based sea ice forecasting system, ECMWF SEAS5, and a statistical benchmark. This dataset includes three types of data from the study. Firstly, IceNet''s SIP forecasts from 2012/1 - 2020/9. Secondly, the 25 neural network files underlying the IceNet model. Thirdly, CSV files of results from the study. The codebase associated with this work includes a script to download this dataset and reproduce all the paper''s figures. This dataset is supported by Wave 1 of The UKRI Strategic Priorities Fund under the EPSRC Grant EP/T001569/1, particularly the "AI for Science" theme within that grant and The Alan Turing Institute. The dataset is also supported by the NERC ACSIS project (grant NE/N018028/1).

The dataset is the output of a statistical model which downscales ERA5 monthly precipitation data using gauge measurements from the Upper Beas and Sutlej Basins in the Western Himalayas. Multi-Fidelity Gaussian Processes (MFGPs) are used to generate more accurate precipitation values between 1980 and 2012, including over ungauged areas of the basins. MFGPs are a probabilistic machine learning method that provides principled uncertainty estimates via the prediction of probability distributions. These predictions can therefore be used to estimate the likelihood of extreme precipitation events which have led to droughts, floods, and landslides. Funding from UK Engineering and Physical Sciences Research Council [grant number: 2270379].

Monitoring whales in remote regions is important for their conservation, using traditional survey platforms (boat and plane) is logistically difficult. The use of very high-resolution satellite imagery to survey whales, particularly in remote regions, is gaining interest and momentum. However, development is hindered by the lack of automated systems to detect whales. Such a system requires an open source library containing examples of whales and confounding features in satellite imagery. Here we present such a database, created by surveying 6,300 km2 of satellite imagery in various regions across the globe, which allowed us to detect 633 whale objects. This dataset contains image chips as png files. Funding was provided from a BAS Innovation Voucher.

Monitoring whales in remote regions is important for their conservation, using traditional survey platforms (boat and plane) is logistically difficult. The use of very high-resolution satellite imagery to survey whales, particularly in remote regions, is gaining interest and momentum. However, development is hindered by the lack of automated systems to detect whales. Such a system requires an open source library containing examples of whales and confounding features in satellite imagery. Here we present such a database, created by surveying 6,300 km2 of satellite imagery in various regions across the globe, which allowed us to detect 633 whale objects and 120 confounding features. Funding was provided from a BAS Innovation Voucher.