A Tile Service-driven Architecture for Online Climate Analysis with an Application to Estimation of Ocean Carbon Flux

The deteriorating environmental system and increasing number of time-varying datasets have generated a significant demand for new techniques to support online high-efficiency climate analysis.

To realize online high-efficiency visualization and analysis, we adopt a pre-processing method for raw data to build a lossless tile set based on the image pyramid structure for each image according to its spatial extent and spatial reference. With the proposed tile service, we can perform online high-efficiency ocean carbon flux estimation by large-sized remote sensing images.As the original image has been broken into lightweight workable tiles, the entire calculation process is applied to each tile in parallel (Figure 1), and the calculation results can be loaded onto the virtual globe directly tile by tile without any merging.

Figure 1 An example of air-sea CO2 flux calculation by multiple remote sensing image data.The whole calculation process is applied to each tile in parallel.

We also compared the loading times of lossless tiles and the original image under different network bandwidths and the results are shown in Figure 2. We only simulate in the way of point-to-point transmission; if we are facing real-world applications insomuch that the upload bandwidths of servers and the download bandwidths of clients are both limited, it will challenge the transmission of original images substantially. In contrast, the tile strategy could behave better at speeding loading times up for the transmission of massive data from the server to the client under a limited network bandwidth.

Figure 2 Comparison of data loading time.(a/b) Original image/lossless tiles loading time comparison between different bandwidths and image counts.(c) Loading time ratio comparison between the original image and the lossless tiles.

Reference: 

Ye, Weiwen, Feng Zhang, Yan Bai, Zhenhong Du, and RenYi Liu. "A tile service-driven architecture for online climate analysis with an application to estimation of ocean carbon flux." Environmental Modelling & Software 118 (2019): 120-133.