python–读取TRMM-3B43月平均降水
绘制气候态空间分布图(陆地区域做掩膜)
TRMM降水数据介绍
热带降雨测量任务(The Tropical Rainfall Measuring Mission,TRMM)
是美国国家航空航天局(NASA
)和日本国家太空发展署(National Space Development Agency
)的一项联合太空任务,旨在监测和研究热带和亚热带降雨及其相关的能量释放
。该任务使用5种仪器: 降水雷达(Precipitation Radar,PR)、 TRMM 微波成像仪(TRMM Microwave Imager,TMI)、可见红外扫描仪(Visible Infrared Scanner,VIRS)、云与地球辐射能量系统(Clouds & Earths Radiant Energy System,CERES)和闪电成像传感器(Lightning Imaging Sensor,LSI)。TMI 和 PR 是用于降水的主要仪器。这些仪器被用于形成 TRMM 多卫星降水分析(TRMM Multi-satellite Precipitation Analysis,TMPA)的 TRMM 组合仪器(TRMM Combined Instrument ,TCI)校准数据集(TRMM 2B31)的算法中,其 TMPA 3B43月平均降水量
和 TMPA 3B42日平均和次日(3小时)平均
是最相关的 TRMM 相关气候研究的产品。3B42和3B43
的空间分辨率为0.25 ° ,1998年至今覆盖北纬50 ° 至南纬50 ° 。
本文中用到的数据主要为TRMM-3B43月平均产品,用于绘制降水的气候态空间分布图
TRMM-3B43产品如下所示:
- 空间分辨率:0.25°
- 时间覆盖范围:1999.01 - 2020.01
- 经纬度范围: 经度:0-360°,纬度:-50°S-50°N
- 单位为:mm/hour
- 降水类型 : 累计降水
这里使用的python系统环境:linux系统,因为windows上好多库都不好用
数据处理
- 这里所使用的3B43降水资料数据为
.HDF
格式,因此需要使用pyhdf
这个库来读取,不习惯的可以下载netcdf
的数据格式 - 由于数据中缺少经纬度信息(也可能是我没有找到),为了实现区域切片,这里手动造了一个dataarray的数据,从而实现切片的过程
- 数据中读取的变量为
precipitation
,读取完之后是个二维的数组,为了给他加上时间纬度,所以手动给他进行了扩维,之后实现多年的气候态平均计算 - 使用
global_land_mask
实现对于陆地区域的掩膜处理 - 使用
cnmaps
实现中国的区域绘制,这里的cnmaps的库在windows上可能不好安装,也直接使用cartopyax.coastlines('50m')
自带的海岸线 - 将原始数据单位转化为 mm/year,这里只是简单的转换
*24*365
代码实现
1、首先读取10年的数据路径
import xarray as xr
import os,glob
import numpy as np
from pyhdf.SD import SD
path = '/Datadisk/TRMM/3B43/'
file_list = []
for year in range(2009,2019):
folder = os.path.join(path, str(year))
file_name = glob.glob(folder+'/3B43.'+str(year)+'*.HDF')
file_name.sort()
file_list.extend(file_name)
file_list.sort()
2、封装数据读取函数,并对需要的区域进行切片,我选择的区域为经度:[90.0, 145],纬度:[-10, 55],并将循环读取的10年月平均数组创建为dataarray的格式方面后续掩膜,这里的时间可以通过pandas自己创建时间序列,我这里偷懒直接读取了之前处理过的月平均gpcp的time了
dt = xr.open_dataset("/gpcp_monthly_mask.nc")
precip = dt.sel(time=slice('2009','2018'),lat =slice(-10,55),lon = slice(90,145)).precip
time_num = precip.time.values
def get_data(path,time):
da = SD(path)
pre = da.select('precipitation')[:]
pre = np.expand_dims(pre,axis=0)
lon = np.arange(-180,180.,0.25)
lat = np.arange(-50,50.,0.25)
time = time
# time = datetime.datetime.utcfromtimestamp(tim).strftime('%Y-%m-%d %H:%M:%S')
da = xr.DataArray(pre,
dims=['time','lon','lat'],
coords=dict(
lon=(['lon'], lon),
lat=(['lat'], lat),
time=(['time'],[time])),
)
##############################################################################
lon_range = [90.0, 145]
lat_range = [-10, 55]
da = da.sel(lon=slice(*lon_range), lat=slice(*lat_range))
x ,y = da.lon,da.lat
return da,x,y
rain = np.zeros((len(file_list),221,240))
for i in range(len(file_list)):
print(i)
da,x,y = get_data(file_list[i], time_num[i])
rain[i] = da
ds = xr.DataArray(rain,
dims=['time','lon','lat'],
coords=dict(
lon=(['lon'], x.data),
lat=(['lat'], y.data),
time=(['time'],time_num)),
)
3、对数据的陆地部分进行掩膜,并计算气候态平均,转换单位为mm/year
from global_land_mask import globe
def mask_land(ds, label='land', lonname='lon'):
if lonname == 'lon':
lat = ds.lat.data
lon = ds.lon.data
if np.any(lon > 180):
lon = lon - 180
lons, lats = np.meshgrid(lon, lat)
mask = globe.is_ocean(lats, lons)
temp = []
temp = mask[:, 0:(len(lon) // 2)].copy()
mask[:, 0:(len(lon) // 2)] = mask[:, (len(lon) // 2):]
mask[:, (len(lon) // 2):] = temp
else:
lons, lats = np.meshgrid(lon, lat)# Make a grid
mask = globe.is_ocean(lats, lons)# Get whether the points are on ocean.
ds.coords['mask'] = (('lat', 'lon'), mask)
elif lonname == 'longitude':
lat = ds.latitude.data
lon = ds.longitude.data
if np.any(lon > 180):
lon = lon - 180
lons, lats = np.meshgrid(lon, lat)
mask = globe.is_ocean(lats, lons)
temp = []
temp = mask[:, 0:(len(lon) // 2)].copy()
mask[:, 0:(len(lon) // 2)] = mask[:, (len(lon) // 2):]
mask[:, (len(lon) // 2):] = temp
else:
lons, lats = np.meshgrid(lon, lat)
mask = globe.is_ocean(lats, lons)
lons, lats = np.meshgrid(lon, lat)
mask = globe.is_ocean(lats, lons)
ds.coords['mask'] = (('latitude', 'longitude'), mask)
if label == 'land':
ds = ds.where(ds.mask == True)
elif label == 'ocean':
ds = ds.where(ds.mask == False)
return ds
data = mask_land(ds,'land')
precip_mean = np.nanmean(data,axis=0)*24*365
4、绘图,保存图片
import cartopy.feature as cfeature
from cartopy.mpl.ticker import LongitudeFormatter,LatitudeFormatter
import cmaps
import matplotlib.pyplot as plt
import cartopy.crs as ccrs
import matplotlib.ticker as mticker
from cartopy.mpl.gridliner import LONGITUDE_FORMATTER, LATITUDE_FORMATTER
from cnmaps import get_adm_maps, draw_maps
box = [100,140,0,50]
xstep,ystep = 10,10
proj = ccrs.PlateCarree(central_longitude=180)
plt.rcParams['font.family'] = 'Times New Roman',
fig = plt.figure(figsize=(8,7),dpi=200)
fig.tight_layout()
ax = fig.add_axes([0.1,0.2,0.8,0.7],projection = proj)
ax.set_extent(box,crs=ccrs.PlateCarree())
draw_maps(get_adm_maps(level='国')) #这里如果库不好安装的话可以使用下面注释的代码,cartopy自带的海岸线
# ax.coastlines('50m')
ax.set_xticks(np.arange(box[0],box[1]+xstep, xstep),crs=ccrs.PlateCarree())
ax.set_yticks(np.arange(box[2], box[3]+1, ystep),crs=ccrs.PlateCarree())
lon_formatter = LongitudeFormatter(zero_direction_label=False)#True/False
lat_formatter = LatitudeFormatter()
ax.xaxis.set_major_formatter(lon_formatter)
ax.yaxis.set_major_formatter(lat_formatter)
ax.set_title('TRMM(mm/year)',fontsize=16,pad=8,loc='left')
ax.tick_params( which='both',direction='in',
width=0.7,
pad=8,
labelsize=14,
bottom=True, left=True, right=True, top=True)
c = ax.contourf(x,y,precip_mean.T,
levels=np.arange(200,3300,100),
extend='both',
transform=ccrs.PlateCarree(),
cmap=cmaps.NCV_jet)
cb=plt.colorbar(c,
shrink=0.98,
orientation='vertical',
aspect=28,
)
cb.ax.tick_params(labelsize=10,which='both',direction='in',)
plt.show()
fig.savefig('./TRMM_10year_monthly.png',dpi=500)
全部代码
# -*- coding: utf-8 -*-
"""
@author: %(jixianpu)s
Email : 211311040008@hhu.edu.cn
introduction : keep learning althongh walk slowly
"""
import xarray as xr
import os,glob
import numpy as np
from pyhdf.SD import SD
path = '/Datadisk/TRMM/3B43/'
file_list = []
for year in range(2009,2019):
folder = os.path.join(path, str(year))
file_name = glob.glob(folder+'/3B43.'+str(year)+'*.HDF')
file_name.sort()
file_list.extend(file_name)
file_list.sort()
dt = xr.open_dataset("/gpcp_monthly_mask.nc")
precip = dt.sel(time=slice('2009','2018'),lat =slice(-10,55),lon = slice(90,145)).precip
time_num = precip.time.values
def get_data(path,time):
da = SD(path)
pre = da.select('precipitation')[:]
pre = np.expand_dims(pre,axis=0)
lon = np.arange(-180,180.,0.25)
lat = np.arange(-50,50.,0.25)
time = time
# time = datetime.datetime.utcfromtimestamp(tim).strftime('%Y-%m-%d %H:%M:%S')
da = xr.DataArray(pre,
dims=['time','lon','lat'],
coords=dict(
lon=(['lon'], lon),
lat=(['lat'], lat),
time=(['time'],[time])),
)
##############################################################################
lon_range = [90.0, 145]
lat_range = [-10, 55]
da = da.sel(lon=slice(*lon_range), lat=slice(*lat_range))
x ,y = da.lon,da.lat
return da,x,y
rain = np.zeros((len(file_list),221,240))
for i in range(len(file_list)):
print(i)
da,x,y = get_data(file_list[i], time_num[i])
rain[i] = da
ds = xr.DataArray(rain,
dims=['time','lon','lat'],
coords=dict(
lon=(['lon'], x.data),
lat=(['lat'], y.data),
time=(['time'],time_num)),
)
from global_land_mask import globe
def mask_land(ds, label='land', lonname='lon'):
if lonname == 'lon':
lat = ds.lat.data
lon = ds.lon.data
if np.any(lon > 180):
lon = lon - 180
lons, lats = np.meshgrid(lon, lat)
mask = globe.is_ocean(lats, lons)
temp = []
temp = mask[:, 0:(len(lon) // 2)].copy()
mask[:, 0:(len(lon) // 2)] = mask[:, (len(lon) // 2):]
mask[:, (len(lon) // 2):] = temp
else:
lons, lats = np.meshgrid(lon, lat)# Make a grid
mask = globe.is_ocean(lats, lons)# Get whether the points are on ocean.
ds.coords['mask'] = (('lat', 'lon'), mask)
elif lonname == 'longitude':
lat = ds.latitude.data
lon = ds.longitude.data
if np.any(lon > 180):
lon = lon - 180
lons, lats = np.meshgrid(lon, lat)
mask = globe.is_ocean(lats, lons)
temp = []
temp = mask[:, 0:(len(lon) // 2)].copy()
mask[:, 0:(len(lon) // 2)] = mask[:, (len(lon) // 2):]
mask[:, (len(lon) // 2):] = temp
else:
lons, lats = np.meshgrid(lon, lat)
mask = globe.is_ocean(lats, lons)
lons, lats = np.meshgrid(lon, lat)
mask = globe.is_ocean(lats, lons)
ds.coords['mask'] = (('latitude', 'longitude'), mask)
if label == 'land':
ds = ds.where(ds.mask == True)
elif label == 'ocean':
ds = ds.where(ds.mask == False)
return ds
data = mask_land(ds,'land')
precip_mean = np.nanmean(data,axis=0)*24*365
import cartopy.feature as cfeature
from cartopy.mpl.ticker import LongitudeFormatter,LatitudeFormatter
import cmaps
import matplotlib.pyplot as plt
import cartopy.crs as ccrs
import matplotlib.ticker as mticker
from cartopy.mpl.gridliner import LONGITUDE_FORMATTER, LATITUDE_FORMATTER
from cnmaps import get_adm_maps, draw_maps
box = [100,140,0,50]
xstep,ystep = 10,10
proj = ccrs.PlateCarree(central_longitude=180)
plt.rcParams['font.family'] = 'Times New Roman',
fig = plt.figure(figsize=(8,7),dpi=200)
fig.tight_layout()
ax = fig.add_axes([0.1,0.2,0.8,0.7],projection = proj)
ax.set_extent(box,crs=ccrs.PlateCarree())
draw_maps(get_adm_maps(level='国'))
# ax.coastlines('50m')
ax.set_xticks(np.arange(box[0],box[1]+xstep, xstep),crs=ccrs.PlateCarree())
ax.set_yticks(np.arange(box[2], box[3]+1, ystep),crs=ccrs.PlateCarree())
lon_formatter = LongitudeFormatter(zero_direction_label=False)#True/False
lat_formatter = LatitudeFormatter()
ax.xaxis.set_major_formatter(lon_formatter)
ax.yaxis.set_major_formatter(lat_formatter)
ax.set_title('TRMM(mm/year)',fontsize=16,pad=8,loc='left')
ax.tick_params( which='both',direction='in',
width=0.7,
pad=8,
labelsize=14,
bottom=True, left=True, right=True, top=True)
c = ax.contourf(x,y,precip_mean.T,
levels=np.arange(200,3300,100),
extend='both',
transform=ccrs.PlateCarree(),
cmap=cmaps.NCV_jet)
cb=plt.colorbar(c,
shrink=0.98,
orientation='vertical',
aspect=28,
)
cb.ax.tick_params(labelsize=10,which='both',direction='in',)
plt.show()
fig.savefig('./TRMM_10year_monthly.png',dpi=500)
引用参考
TRMM: Tropical Rainfall Measuring Mission https://climatedataguide.ucar.edu/climate-data/trmm-tropical-rainfall-measuring-mission
Monthly 0.25° x 0.25° TRMM multi-satellite and Other Sources Rainfall (3B43) http://apdrc.soest.hawaii.edu/datadoc/trmm_3b43.php
TRMM 3B43: Monthly Precipitation Estimates https://developers.google.com/earth-engine/datasets/catalog/TRMM_3B43V7
中巴经济走廊TRMM_3B43月降水数据(1998-2017年):http://www.ncdc.ac.cn/portal/metadata/4b9504fa-0e34-47c9-a755-91d3f3253312
TRMM (TMPA/3B43) Rainfall Estimate L3 1 month 0.25 degree x 0.25 degree V7 (TRMM_3B43)(GES 官网介绍):https://disc.gsfc.nasa.gov/datasets/TRMM_3B43_7/summary
国家海洋遥感在线分析平台 https://www.satco2.com/index.php?m=content&c=index&a=show&catid=317&id=217