search

jeonghwan723 / DL_ENSO

CNN for climate forecast
MIT License
80 stars 36 forks source link

data processing #13

Open lzfzl opened 2 months ago

lzfzl commented 2 months ago

Hi, Excellent work! I’m trying to reproduce the results of your paper and I get a similiar result using the data provided by you. But I can’t get the same result when I use the data processed by myself, and I find difference between your data and mine.There may be something wrong during my processing, could you please tell me how you process the original data? Thanks. lzfzl

jeonghwan723 commented 1 month ago

Dear @lzfzl,

Thank you for your interest in our work. We preprocessed the inputs in the following way:

  1. We horizontally interpolated the sea surface temperature and oceanic heat content data at 5 x 5 degree.
  2. Removed the monthly climatology from total fields.
  3. All masked values (i.e., land values) have been replaced with zeros.

I hope this is helpful to you. Jeong-Hwan

lzfzl commented 1 month ago

@jeonghwan723 Thank you so much for your help, I forget to replace the masked values with zeros at the beginning and let me have a try. By the way, I interpreting "oceanic heat content (HC, in units of °C)" in the paper as "Depth Average Potential Temperature of Upper 300m", is that right?

lzfzl

jeonghwan723 commented 1 month ago

@lzfzl That's right :)

Jeong-Hwan