Reanalysis of sea surface temperature for the PDO forecast using the CNN model

Citation Author(s):
Nutta
Skanupong
Institute of Oceanology Chinese Academy of Sciences
Yongsheng
Xu
Institute of Oceanology Chinese Academy of Sciences
Lejiang
Yu
Polar Research Institute of China
Zhang
Wan
Nanyang Technological University
Shuo
Wang
University of Birmingham
Submitted by:
Nutta Skanupong
Last updated:
Fri, 03/29/2024 - 05:32
DOI:
10.21227/n2qg-v223
Data Format:
License:
0
0 ratings - Please login to submit your rating.

Abstract 

This data is the reanalysis of sea surface temperature provided by Extended Reconstruction Sea Surface Temperature version 5 (ERSST v.5) from January 1854 to December 2022, Hadley Centre Global Sea Ice and Sea Surface Temperature (HadISST) from January 1870 to December 2022, and COBE-SST2 Sea Surface Temperature and Ice (COBE-SST2) from January 1854 to December 2022. All data is re-gridded to have the same spatial resolution of 2.0° × 2.0°, and the grid spans from 88°N to 88°S and 0°E to 358°E via bilinear interpolation from the initial grid. This dataset is in NetCDF4 format. The Convolutional Neural Network (CNN) model applies this data to forecast the Pacific Decadal Oscillation (PDO).

Instructions: 

The dataset folder comprises three datasets.

The first data is the reanalysis of Extended Reconstruction Sea Surface Temperature version 5 (ERSST v.5), which offers a monthly global spatial complete record of sea surface temperature in degrees Celsius (°C) from January 1854 to December 2022.

The second data is the reanalysis of Hadley Centre Global Sea Ice and Sea Surface Temperature (HadISST), which combines monthly measurements of sea surface temperature and sea ice concentration. The initial dataset is presented in a global grid format with a resolution of 1.0° × 1.0°. This analysis data is modified to have a spatial resolution of 2.0° × 2.0°, and the grid spans from 88°N to 88°S and 0°E to 358°E. This provided data only focuses on the sea surface temperature in °C from January 1870 to December 2022.

The third data is the reanalysis of COBE-SST2 Sea Surface Temperature and Ice (COBE-SST2), which is constructed from a daily sea surface temperature by combining a trend, interannual variations, and daily changes. The initial dataset is presented in a global grid format with a resolution of 1.0° × 1.0°, and the grid spans from 89.5 °N to 89.5 °S and 0.5 °E to 359.5 °E. This sea surface temperature data has been adjusted to a spatial resolution of 2.0° × 2.0°, and the grid covers a range from 88°N to 88°S and 0°E to 358°E, spanning from January 1854 to December 2022. The temperature is measured in °C.

Funding Agency: 
Laoshan Laboratory science and technology innovation projects
Grant Number: 
U22A20587