Lossy compression of Earth system model data based on a hierarchical tensor with Adaptive-HGFDR (v1.0)

Abstract. Lossy compression has been applied to the data of large-scale Earth system model (ESMD) due its advantages a high ratio. However, few lossy methods consider both global and local multidimensional coupling correlations, which could lead information loss in approximation compression. Here, an adaptive method, hierarchical geospatial field representation (Adaptive-HGFDR), is developed based on foundation stream method for called blocked (Blocked-HGFDR). In addition, original Blocked-HGFDR also improved from following perspectives. Firstly, are divided into series blocks more balanced size reduce effect dimensional unbalance ESMD. Following this, mathematical relationship between parameter error Blocked-HGFDR, control mechanism determine optimal given error. By assigning each block independent parameter, Adaptive-HGFDR can capture variation correlations improve accuracy. Experiments carried out Community System Model (CESM) data. The results show that our higher ratio uniform distributions compared with ZFP Blocked-HGFDR. For among 22 climate variables, achieve good performances most flux variables significant spatiotemporal heterogeneity fast changing rate. This study provides new potential