Skip to content

Maps

Density Maps

Maps

ScreenshotScreenshot

ScreenshotScreenshot

ScreenshotScreenshot

FIGURE: left: Probability density maps for exceeding the 99th percentile of daily precipitation in Jordan within 1981 to 2010 applied to date. right: Probability density maps for record breaking daily precipitation events from from 1981 to date. Three different gridded data products are used (UERRA, ERA5, CHIRPS).


Code

Importing

import matplotlib
matplotlib.use('Agg') # Must be before importing matplotlib.pyplot or pylab!
import matplotlib.pyplot as plt
import matplotlib.pyplot as P
from mpl_toolkits.basemap import Basemap,shiftgrid,maskoceans
from matplotlib.path import Path
from matplotlib.patches import PathPatch
import shapefile
import numpy as N
from netCDF4 import Dataset, num2date,date2num
import datetime
from scipy import signal,stats
import matplotlib.colors as colors
from matplotlib.offsetbox import AnchoredText
import matplotlib as mpl


plt.style.use('seaborn-talk')

params = {      'legend.fontsize': 8,\
                'font.family': 'serif',\
                }
plt.rcParams.update(params)

Reading

city = {

   'Amman':[35.930359,31.963158],
   'Aqaba':[35.00778,29.52667],
   'WadiMusa':[35.480125,30.321635],
}


datas = ['UERRA','ERA5','CHIRPS']
modes = ['THRESHOLD','RECORD']

for data in datas:
    for mode in modes:

        if(data=='UERRA'): file = '../../data/ClimateExplorer/tp_daily_uerra_19810101-20181231.cdf'
        if(data=='CHIRPS'): file = '../../data/ClimateExplorer/v2p0chirps_25_34-40E_29-34N.cdf'
        if(data=='ERA5'): file = '../../data/ClimateExplorer/era5_tp_daily_af_34-40E_29-34N_su.cdf'

        nc = Dataset(file,'r')
        lon = N.array(nc.variables['longitude'][:])
        lat = N.array(nc.variables['latitude'][:])
        dat = N.array(nc.variables['tp'][:])
        tim = nc.variables['time']

        print (dat.shape,N.max(dat),N.min(dat))

        dat[dat>1000] = N.nan

        tim = num2date(tim[:],units=tim.units,calendar=tim.calendar) 

        jj = []
        mm = []
        dd = []

        for it in tim:

            jj.append(it.year)
            mm.append(it.month)
            dd.append(it.day)

        jj = N.array(jj)
        mm = N.array(mm)
        dd = N.array(dd)

        nc.close()

        nd = dat.shape[0]
        ny = dat.shape[1]
        nx = dat.shape[2]

        id = N.where((jj>=1981)&(jj<=2010))[0]

        tmp = N.nanmax(N.reshape(dat,(nd,ny*nx)),1)
        rx = N.percentile(tmp[id],95)

        print (data,mode,rx)

        max = N.nanmax(dat[id,:,:],0)

        xx = []
        yy = []

        for y in range(ny):
            for x in range(nx):
                for d in range(nd):

                   if(mode=='THRESHOLD'): 

                        if((dat[d,y,x]>=rx)&(dat[d,y,x]<1000.)):

                            xx.append(lon[y,x])
                            yy.append(lat[y,x])

                   if(mode=='RECORD'): 

                        if((dat[d,y,x]>=max[y,x])&(dat[d,y,x]<1000.)):

                            max[y,x] = dat[d,y,x]

                            xx.append(lon[y,x])
                            yy.append(lat[y,x])


        xx = N.array(xx)
        yy = N.array(yy)

        values = N.vstack([xx,yy])
        kernel = stats.gaussian_kde(values,bw_method='silverman')

        X,Y = N.mgrid[34:40:100j,29:34:100j]

        positions = N.vstack([X.ravel(),Y.ravel()])
        Z2 = N.reshape(kernel(positions).T, X.shape)    

        fig = P.figure(figsize=(8,4))
        ax = P.subplot(111)

        m = Basemap(projection='cyl',llcrnrlat=29,urcrnrlat=33.5,llcrnrlon=34.5,urcrnrlon=39.5,resolution='h')
        #m.drawcountries()
        #m.drawcoastlines() 

        xx,yy = m(X,Y)

        lev = N.arange(0.1,1.0,0.1)

        cl = m.contourf(xx,yy,Z2,lev,cmap=P.get_cmap('terrain_r'),extend='both',zorder=4)
        cs = m.contour(xx,yy,Z2,lev,colors='k',linestyles='solid',zorder=5,linewidths=0.5)
        #P.clabel(cs,cs.levels,inline=True,fmt='%2.1f', fontsize=8)

        jmin = N.min(jj)
        jmax = N.max(jj)

        at = AnchoredText('%s %i-%i: %s 95th = %.1f mm/d'%(data,jmin,jmax,mode,rx),prop=dict(size=7,weight='bold'),frameon=True,loc='upper left')
        at.patch.set_boxstyle("round,pad=0.,rounding_size=0.1")
        ax.add_artist(at)

        #at = AnchoredText('95th = %.1f mm/d'%(rx),prop=dict(size=6,weight='bold'),frameon=True,loc='lower left')
        #at.patch.set_boxstyle("round,pad=0.,rounding_size=0.1")
        #ax.add_artist(at)

        for c in city:

            x,y = m(city[c][0],city[c][1])

            m.scatter(x,y,s=30,c='r',ec='k',lw=0.5,zorder=9)
            plt.text(x,y,c,fontsize=8,weight='bold',zorder=10)    

        sf = shapefile.Reader("../../data/shp/JOR_adm0")

        for shape_rec in sf.shapeRecords():
            if shape_rec.record[1] == 'JOR':
                vertices = []
                codes = []
                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(m(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,fill=False,lw=0.5)

        for contour in cs.collections:
            contour.set_clip_path(clip)
            ax.add_patch(clip)

        for contour in cl.collections:
            contour.set_clip_path(clip)
            ax.add_patch(clip)

        ax.axis('off')

        ax1 = fig.add_axes([0.64, 0.15, 0.02, 0.35])

        cb = mpl.colorbar.ColorbarBase(ax1,cmap=plt.get_cmap('terrain_r'),spacing='uniform',orientation='vertical',boundaries=lev,extend='both')
        cb.set_label('Prob.',labelpad=-30,y=1.2,rotation=0,fontsize=10,weight='bold')
        cb.ax.tick_params(labelsize=10)
        cb.solids.set_rasterized(True)
        cb.solids.set_edgecolor("k")

        plt.savefig('./img/%s_%s.png'%(data,mode),dpi=240,transparent=False,bbox_inches='tight',pad_inches=0.0)