A note on precision-preserving compression of scientific data

Abstract. Lossy compression of scientific data arrays is a powerful tool to save network bandwidth and storage space. Properly applied lossy compression can reduce the size of a dataset by orders of magnitude keeping all essential information, whereas a wrong choice of lossy compression parameters leads to the loss of valuable data. The paper considers statistical properties of several lossy compression methods implemented in "NetCDF operators" (NCO), a popular tool for handling and transformation of numerical data in NetCDF format. We compare the effects of imprecisions and artifacts resulting from use of a lossy compression of floating-point data arrays. In particular, we show that a popular Bit Grooming algorithm (default in NCO) has sub-optimal accuracy and produces substantial artifacts in multipoint statistics. We suggest a simple implementation of two algorithms that are free from these artifacts and have twice higher precision. Besides that, we suggest a way to rectify the data already processed with Bit Grooming. The algorithm has been contributed to NCO mainstream. The supplementary material contains the implementation of the algorithm in Python 3.