概述（前言）
本文作为上一篇博客《MATLAB处理GPM(IMERG/GSMaP)卫星降水数据》的后续，作者选择另一种语言python来实现处理降水领域高分辨率降水数据GPM-IMERG和GPM-GSMaP，两种数据的详细介绍、下载方式以及matlab读取参见上一篇文章《MATLAB处理GPM(IMERG/GSMaP)卫星降水数据》：
image.png
正文
Python 处理GPM-IMERG数据
1. 数据下载(python ftplib)
参见文章《python 实现FTP文件批量下载》
2.数据读取(python)
#!/usr/bin/env python
# -*- coding: utf-8 -*-
# @Time : 2019/2/28 17:28
# @Author : zengsk in HoHai
# @File : Calc_AVG_IMERG.py
'''
Calculating the average precipitation over a given region
'''
# Import packages
import os
import h5py
import numpy as np
import pandas as pd
import warnings
warnings.simplefilter("ignore")
# header
header = '''ncols 3600
nrows 1800
xllcorner -180
yllcorner -90
cellsize 0.1
NODATA_value -999.0'''
# dataset directory
sPath = r'D:\降水数据\satellite_pre\IMERG\late\201508'
# mask region
mask = pd.read_table(r'.\assets\china_mask_gpm_01.txt', sep='\s+', header=None, skiprows=6).values
rain = np.empty([1800, 3600], dtype=float)
tag = np.empty(rain.shape, dtype=float) # Record the effective precip times per grid
Files = os.listdir(sPath)
for file in Files:
fpath = os.path.join(sPath, file)
if os.path.splitext(fpath)[1] == '.RT-H5':
print(os.path.basename(fpath))
f = h5py.File(fpath) # Open the H5 file, return the File class
Uncal = f['/Grid/precipitationUncal'].value
Uncal = np.transpose(Uncal) # size: 1800 x 3600
Uncal = np.flipud(Uncal) # convert to 90N ~ 90S
tag[Uncal>=0] = tag[Uncal>=0] + 1
Uncal[Uncal<0] = 0
Uncal /= 2
rain += Uncal
# output
rainfall_avg = rain / tag * 2 * 24 # unit conversion
rainfall_avg[np.isnan(rainfall_avg)] = -999.00 # NODATA_value
rainfall_avg[mask<0] = -999.00
ofname = r'.\output\IMERG_AVG_201508mon.txt'
np.savetxt(ofname, rainfall_avg, fmt='%7.2f ', comments='', header=header)
print("\n***************** 数据处理完成!!! Nice Job!!! ****************")
Python 处理GPM-GSMaP数据
1. 数据下载(python ftplib)
参见文章《python 实现FTP文件批量下载》
2.数据读取(python)
#!/usr/bin/env python
# -*- coding: utf-8 -*-
# @Time : 2019/4/13 18:34
# @Author : zengsk in HoHai
# @File : Calc_Avg_GSMaP.py
'''
Calculating the average precipitation over a given region
GsMaP info:
60N-->60S, 0-->360E;
0.1*0.1degree;
1 hourly or daily;
4-byte-float;
'''
import os
import struct
import numpy as np
import pandas as pd
import warnings
warnings.simplefilter("ignore")
# header
header = '''ncols 3600
nrows 1200
xllcorner 0
yllcorner -60
cellsize 0.1
NODATA_value -999.0'''
# dataset directory
sPath = r'D:\data\GSMaP'
# mask region file
mask = pd.read_table(r'.\assets\china_mask_gpm_01.txt', sep='\s+', header=None, skiprows=6).values
rain = np.empty([1200, 3600], dtype=float)
tag = np.empty(rain.shape, dtype=float) # Record the effective precip times per grid
Files = os.listdir(sPath)
for file in Files:
filepath = os.path.join(sPath, file)
if '.dat' == os.path.splitext(filepath)[1]:
print(filepath)
with open(filepath, 'rb') as fid:
data = fid.read()
data = struct.unpack('f'*3600*1200, data)
data = np.array(data)
data.resize(1200,3600)
tag[data >= 0] = tag[data >= 0] + 1
data[data < 0] = 0.0
data *= 24 # unit conversion
rain += data
# output
rainfall_avg = rain / tag
rainfall_avg[np.isnan(rainfall_avg)] = -999.00 # NODATA_value
rainfall_avg[mask<0] = -999.00
ofname = r'.\output\201607avg_gsmap.txt'
np.savetxt(ofname, rainfall_avg, fmt='%7.2f ', comments='', header=header)
print("\n+---------------- 数据处理完成!!! Nice Job!!! -----------------+\n")
最后来一张IMERG 2015年08月全球平均降水量分布图(由上述代码读取)
IMERG 2015年08月全球平均降水量分布
