ERA5数据批量下载-多进程

本人之前的博客中介绍了批量下载ERA5数据的方法，在这里再此对方法进行优化，采用多进程下载。采用的Python 的Multiprocessing。

可以根据需求选择相应的参数、空间范围、分辨率、时间等等。

__author__ = "LNH"
import cdsapi
import numpy as np
import calendar
import os  
from multiprocessing import Process

def download(c, var, year, month, day, h, area):
    m = '{:0>2}'.format(str(month))
    d = '{:0>2}'.format(str(day))
    
    c.retrieve(
        'reanalysis-era5-single-levels',
        {
            'variable': var,
            'product_type':'reanalysis',
            'year': year,
            'month': month,
            'day': day,
            'time': h,
            # 'pressure_level': ['1','2','3','5',
            #                     '7','10','20','30',
            #                     '50','70','100','125',
            #                     '150','175','200','225',
            #                     '250','300','350','400',
            #                     '450','500','550','600',
            #                     '650','700','750','775',
            #                     '800','825','850','875',
            #                     '900','925','950','975','1000'],             
            'area': area ,
            'format':'netcdf',
            'grid': '0.25/0.25'
            },
        'era5_' + '_single_levels_' + year + m + d + h[0:2] + '.nc')
    
    return 0
               
if __name__ == '__main__':
# var: 你要下载的变量
# url: 个人账号的url
# key: 密钥
# yearstart, yearend: 数据起止年份
# monthstart， monthend: 数据起止月份
# daystart, dayend: 数据起止天
# h: 小时（默认24小时全部下载，从00:00开始）
# area: 数据的区域 [lat2, lon1, lat1, lon2]
# grid: 数据的分辨率, eg. "0.25/0.25"

        # define directory in which data shall be stored
    os.chdir("F:/")
    
    #账号的url\uid\key
    c = cdsapi.Client(url= "XXXXXXXXXXXXXXXXXXXXXXXXX",
                  key = "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXX")
    
    # c = cdsapi.Client(url= "https://cds.climate.copernicus.eu/api/v2",
    #                   key ")

    
    var = ['10m_u_component_of_wind',
       '10m_v_component_of_wind', 
       'Total_cloud_cover',
       'Total_column_ozone',
       'Total_column_water',
       'Total_column_water_vapour',
       '2m_temperature',
       'Skin_temperature',
       'Sea_surface_temperature',
       'Surface_pressure']
    
    var1 = ['fraction_of_cloud_cover',
        'specific_cloud_liquid_water_content',
        'Ozone_mass_mixing_ratio',
        'Specific_humidity',
        'Temperature',
        'u_component_of_wind',
        'v_component_of_wind',
        'geopotential']
    
    var2 = ["orography"]
    # define the years you want to download
    yearstart = 2003
    yearend = 2003
    # define the start and end month you want to download
    monthstart = 6
    monthend = 6
    # define the start and end day you want to download
    daystart = 7
    dayend = 31
    # define spatial limits of download (eg. around Austria)
    lon1 = 100.0
    lon2 = 150.0
    lat1 = 0.0
    lat2 = 50.0
    
    # create lists
    years = np.array(range(yearstart,yearend+1),dtype="str")
    area = [lat2, lon1, lat1, lon2]
                  
    
    for year in years:
        if (int(year)==yearstart) and (int(year)==yearend):
            months = np.array(range(monthstart,monthend+1),dtype="str")
        elif (year == yearstart) :
            months = np.array(range(monthstart,13),dtype="str")
        elif (year == yearend):
            months = np.array(range(1,monthend + 1),dtype="str")
        else:
            months = np.array(range(1,13),dtype="str")
               
    for month in months:
        m = '{:0>2}'.format(str(month))
        # if int(month) < 10:
        #     m = '0' + month
        # else:
        #     m = month
        
        if(int(year) == yearstart) and (int(year) == yearend) and (int(month) == monthstart) and (int(month) == monthend):
                days = list(np.array(range(daystart,dayend+1),dtype="str"))
        elif (int(year) == yearstart) and (int(month) == monthstart):
            days = list(np.array(range(daystart,calendar.monthrange(int(year),int(month))[1]+1),dtype="str"))
        elif (int(year) == yearend) and (int(month) == monthend):
            days = list(np.array(range(1,dayend+1),dtype="str"))
        else:
            days = list(np.array(range(1,calendar.monthrange(int(year),int(month))[1]+1),dtype="str"))
        
        for day in days:
            d = '{:0>2}'.format(str(day))
            # if int(day) < 10:
            #     d = '0' + day
            # else:
            #     d = day
            
            pro_list = []
            # 24 hours start at 00:00
            
            for hour in range(24):
                h = '{:0>2}'.format(str(hour)) + ":00"
                
                p = Process(target=download, args=(c, var, year, month, day, h, area, ))
                p.start()
                pro_list.append(p)
                
            for p in pro_list:
                p.join()
                print("主进程结束!")
        del p
                

如有疑问欢迎留言 ！