Koistinen-Saltikoff

The methodology proposed by Koistinen and Saltikoff (1998) provides an empirical formula to calculate the probability of precipitation type using temperature and relative humidity observations. Formally, the formula calculates the probability of rain and two thresholds are set to discriminate between snow, sleet and rain. In our case, the equation is flipped, so probability of snow is determined by (1) which may be expressed as

\(\begin{equation*} p(snow) = 1 - \dfrac{1}{1 + e^{22 - 2.7\cdot T - 0.2\cdot RH}} \end{equation*}\)

where T corresponds to temperature in Celsius and RH to relative humidity in %. If p(snow) obtained values are below 0.33 precipitation is in form of rain, if they are between 0.33 and 0.66 in form of sleet and classified as snow if they are above 0.66.

In the following example we’ll show how PyPROS classifies precipitation considering the Koistinen-Saltikoff methodology.

First of all, we’ll import the required libraries.

from pypros.pros import PyPros

As an example, we’ll get the precipitation type classification from different methodologies for Catalonia on 2017-03-25 00.30 UTC. For this purpose we’ll use an air temperature, dew point temperature, digital elevation model (DEM) and reflectivity fields.

Those fields can be found in notebooks/data directory and we’ll keep the path for all of them:

tair_file = '../sample-data/INT_TAIR_20170325_0030.tif'
tdew_file = '../sample-data/INT_TDEW_20170325_0030.tif'
dem_file = '../sample-data/DEM_CAT.tif'

Now, we’ll define those parameters that PyPros class uses and are the same whether the methodology changes or not. These parameters are: variables_files and data_format. For more information on this class, see PyPros Class notebook.

variables_files = [tair_file,
                   tdew_file,
                   dem_file]
data_format = {'vars_files':['tair', 'tdew', 'dem']}

Since we want to apply the Koistinen-Saltikoff methodology, first we’ll define method PyPros parameter as 'ks' and then we’ll set the threshold parameter to None.

method = 'ks'
threshold = None

Now, we’re ready to call PyPros class!

ks = PyPros(variables_files, method, threshold, data_format)

We can get a quicklook of the obtained field using plot_pros function:

import matplotlib.pyplot as plt
plt.imshow(ks.result)
plt.show()

In addition, we can save the precipitation type field in a raster file using save_file function:

ks.save_file(ks.result, '../sample-data/output/ks.tif')

If we have a reflectivity field, we can also apply it as a mask by using refl_mask function and save it as a raster file. However, we’ll have to read first the reflectivity field. For this purpose we need to import gdal.

from osgeo import gdal

refl_file = '../sample-data/CAPPI_XRAD_20170325_0030.tif'
refl_array = gdal.Open(refl_file).ReadAsArray()

Once we’ve read the refl_field we can call the refl_mask function.

ks_masked = ks.refl_mask(refl_array)

ks.save_file(ks_masked, '../sample-data/output/ks_masked.tif')