Use of fill algorithm to merge disputed territories in water demand h…

…istoric maps. Extended MSWX dataset to 1975

src/lisfloodutilities/water-demand-historic/step2_domestic_demand.py

@@ -340,7 +340,10 @@ def main():
         # Load MSWX air temperature data
         mswx_monthly = np.zeros((mapsize_global[0],mapsize_global[1],12),dtype=np.single)*np.NaN
         for month in np.arange(1,13):
-            dset = Dataset(os.path.join(config['mswx_folder'],'Past','Temp','Monthly',str(year)+str(month).zfill(2)+'.nc'))
+            if year<1979:  #MSWX dataset starts from 1979
+                dset = Dataset(os.path.join(config['mswx_folder'],'Past','Temp','Monthly',str(1979)+str(month).zfill(2)+'.nc'))
+            else:
+                dset = Dataset(os.path.join(config['mswx_folder'],'Past','Temp','Monthly',str(year)+str(month).zfill(2)+'.nc'))
             data = np.squeeze(np.array(dset.variables['air_temperature']))
             mswx_monthly[:,:,month-1] = imresize_mean(data,mapsize_global)
         mswx_avg = np.mean(mswx_monthly,axis=2)

src/lisfloodutilities/water-demand-historic/step3a_industrial_demand.py

@@ -731,7 +731,10 @@ def main():
             ndays = monthrange(year,month)[1]
             mswx_month = np.zeros((1800,3600,ndays),dtype=np.single)*np.NaN
             for day in np.arange(1,ndays+1):
-                filepath = os.path.join(config['mswx_folder'],'Past','Temp','Daily',datetime(year,month,day).strftime('%Y%j')+'.nc')
+                if year<1979:   #MSWX dataset starts from 1979
+                    filepath = os.path.join(config['mswx_folder'],'Past','Temp','Daily',datetime(1979,month,day).strftime('%Y%j')+'.nc')
+                else:
+                    filepath = os.path.join(config['mswx_folder'],'Past','Temp','Daily',datetime(year,month,day).strftime('%Y%j')+'.nc')
                 if os.path.isfile(filepath):
                     dset = Dataset(filepath)
                     mswx_month[:,:,day-1] = np.squeeze(np.array(dset.variables['air_temperature']))

src/lisfloodutilities/water-demand-historic/tools.py

@@ -35,8 +35,8 @@ def load_country_code_map(filepath,mapsize):
     assert (country_code_map_refmatrix[0]==-180) & (country_code_map_refmatrix[3]==90)
 
     country_code_map = resize(country_code_map,mapsize,order=0,mode='edge',anti_aliasing=False)
-    country_code_map[country_code_map==158] = 156 # Add Taiwan to China
-    country_code_map[country_code_map==736] = 729 # South Sudan missing from map
+    # apply the fill algorithm to fill the gaps of the Disputed Territories (code>1000), so that the fill algorithm will set each pixel value to their nearest neighbour
+    country_code_map = fill(country_code_map, country_code_map>1000) # CR: check here: the fill algorithm includes also zero as a neighbour country code, so that isles and some land area will be replaced by ocean
     country_code_map[country_code_map==0] = np.NaN
 
     return country_code_map