Python 天气数据处理|克里金插值和可视化
最编程
2024-04-07 13:00:38
...
克里金法(Kriging) 是依据协方差函数对随机过程/随机场进行空间建模和预测(插值)的回归算法。在特定的随机过程,例如固有平稳过程中,克里金法能够给出最优线性无偏估计(Best Linear Unbiased Prediction, BLUP),因此在地统计学中也被称为空间最优无偏估计器(spatial BLUP)。
1、安装模块
!pip install PyKrige
!pip install plotnine
!pip install openpyxl
2、导入模块
import pandas as pd
from pykrige.ok import OrdinaryKriging
import plotnine
from plotnine import *
import geopandas as gpd
import shapefile
import matplotlib.pyplot as plt
import cartopy.crs as ccrs
import cartopy.io.shapereader as shpreader
import cmaps
from matplotlib.path import Path
from matplotlib.patches import PathPatch
3、读取数据
df = pd.read_excel('/home/mw/input/meiyu6520/meiyu_sh_2020.xlsx')
df
# 读取站点经度
lons = df['lon']
# 读取站点纬度
lats = df['lat']
# 读取梅雨量数据
data = df['meiyu']
# 生成经纬度网格点
grid_lon = np.linspace(120.8, 122.1,1300)
grid_lat = np.linspace(30.6, 31.9,1300)
4、克里金(Kriging)插值
OK = OrdinaryKriging(lons, lats, data, variogram_model='gaussian',nlags=6)
z1, ss1 = OK.execute('grid', grid_lon, grid_lat)
z1.shape
输出:
(1300, 1300)
转换成网格
xgrid, ygrid = np.meshgrid(grid_lon, grid_lat)
将插值网格数据整理
df_grid = pd.DataFrame(dict(long=xgrid.flatten(),lat=ygrid.flatten()))
添加插值结果
df_grid["Krig_gaussian"] = z1.flatten()
df_grid
输出:
long lat Krig_gaussian
0 121.000000 30.5 541.076134
1 121.020408 30.5 540.997891
2 121.040816 30.5 540.770766
3 121.061224 30.5 540.381399
4 121.081633 30.5 539.818047
... ... ... ...
2495 121.918367 31.5 567.391506
2496 121.938776 31.5 562.730899
2497 121.959184 31.5 558.418169
2498 121.979592 31.5 554.460195
2499 122.000000 31.5 550.857796
2500 rows × 3 columns
5、读取上海的行政区划
sh = gpd.read_file('/home/mw/input/meiyu6520/Shanghai.shp')
sh
输出:
City District Province Code geometry
0 上海市 普陀区 上海市 310107 POLYGON ((121.35622 31.23362, 121.35418 31.237...
1 上海市 宝山区 上海市 230506 POLYGON ((121.48552 31.31156, 121.48541 31.311...
2 上海市 崇明区 上海市 310151 MULTIPOLYGON (((121.87022 31.29554, 121.86596 ...
3 上海市 奉贤区 上海市 310120 POLYGON ((121.56443 30.99643, 121.57047 30.998...
4 上海市 虹口区 上海市 310109 POLYGON ((121.46828 31.31520, 121.46831 31.316...
5 上海市 黄浦区 上海市 310101 POLYGON ((121.48781 31.24419, 121.49483 31.242...
6 上海市 嘉定区 上海市 310114 POLYGON ((121.33689 31.29506, 121.33650 31.294...
7 上海市 金山区 上海市 310116 POLYGON ((121.00206 30.95104, 121.00764 30.947...
8 上海市 静安区 上海市 310106 POLYGON ((121.46808 31.32032, 121.46809 31.320...
9 上海市 闵行区 上海市 310112 POLYGON ((121.33689 31.23674, 121.33835 31.237...
10 上海市 浦东新区 上海市 310115 MULTIPOLYGON (((121.97077 31.15756, 121.96568 ...
11 上海市 青浦区 上海市 310118 POLYGON ((121.31900 31.15867, 121.31953 31.157...
12 上海市 松江区 上海市 310117 POLYGON ((121.02906 30.94388, 121.02804 30.943...
13 上海市 徐汇区 上海市 310104 POLYGON ((121.45800 31.21929, 121.45807 31.219...
14 上海市 杨浦区 上海市 310110 POLYGON ((121.52288 31.34289, 121.52549 31.346...
15 上海市 长宁区 上海市 310105 POLYGON ((121.43938 31.21444, 121.43946 31.214...
绘图:
sh.plot()
6、插值结果可视化
fig = plt.figure(figsize=(12, 8))
ax = fig.add_subplot(1, 1, 1, projection=ccrs.PlateCarree())
ax.set_extent([120.8, 122.1, 30.6, 31.9], crs=ccrs.PlateCarree())
province = shpreader.Reader('/home/mw/input/meiyu6520/Shanghai.shp')
ax.add_geometries(province.geometries(), crs=ccrs.PlateCarree(), linewidths=0.5,edgecolor='k',facecolor='none')
cf = ax.contourf(xgrid, ygrid, z1, levels=np.linspace(0,800,21), cmap=cmaps.MPL_rainbow, transform=ccrs.PlateCarree())
def shp2clip(originfig, ax, shpfile):
sf = shapefile.Reader(shpfile)
vertices = []
codes = []
for shape_rec in sf.shapeRecords():
pts = shape_rec.shape.points
prt = list(shape_rec.shape.parts) + [len(pts)]
for i in range(len(prt) - 1):
for j in range(prt[i], prt[i + 1]):
vertices.append((pts[j][0], pts[j][1]))
codes += [Path.MOVETO]
codes += [Path.LINETO] * (prt[i + 1] - prt[i] - 2)
codes += [Path.CLOSEPOLY]
clip = Path(vertices, codes)
clip = PathPatch(clip, transform=ax.transData)
for contour in originfig.collections:
contour.set_clip_path(clip)
return contour
shp2clip(cf, ax, '/home/mw/input/meiyu6520/Shanghai.shp')
cb = plt.colorbar(cf)
cb.set_label('Rainfall(mm)',fontsize=15)
plt.show()