[Hpgl-support-eng] Ordinary Kriging on weather data - Reg. urgent help needed!

[sonal a. <son...@gm...>]

Hi,

I am trying to interpolate weather data (temperature) to create contours.
I tried to use ordinary kriging of HPGL. But the results are not
satisfactory.

Details of my codes is as follows :

1. stn_lat.txt, stn_lon.txt, stn_temp.txt  : consists of latitudes,
longitudes and temperature value respectively. [Just for testing purpose i
kept it separately.]

2. one.txt consist of 1's and 0's for the masked value.

3. The latitude,longitude values are basically non-gridded data. Latitude
/longitudes are not equally-spaced. I have Temperature value for all. But
just for testing purpose i replaced some values by 99. and corresponding
mask in one.txt is made 0 instead of 1.

4. the temperature values obtained by doing ordinary kriging is not matching
the original values.What  can be the possible reason??

5. If i want to get temperature values for a grid. how do I do this??


Please help.



My codes are as follows:

*import geo_bsd
from geo_bsd import *
grid = SugarboxGrid(3,1,659)
cov1 = CovarianceModel(type=0, ranges = (5, 5, 5), sill = 1)
import numpy as np
import array
li=[]
inp = open ("input/stn_lat.txt","r")
#read line into array
for line in inp.readlines():
    for i in line.split():
        li.append(float(i))

li2=[]
inp2 = open ("input/stn_lon.txt","r")
#read line into array
for line in inp2.readlines():
    for i in line.split():
        li2.append(float(i))

li4=[]
inp4 = open ("input/stn_temp.txt","r")
#read line into array
for line in inp4.readlines():
    for i in line.split():
        li4.append(float(i))

li3=[]
inp3 = open ("input/one.txt","r")
#read line into array
for line in inp3.readlines():
    for i in line.split():
        li3.append(float(i))

array_prop=np.array(([li2],[li],[li4]),dtype='float32')
array_mask=np.array(([li2],[li],[li3]),dtype='float32')

# define grid.
cont_property = ContProperty(array_prop, array_mask)

prop_result = ordinary_kriging(cont_property, grid, radiuses = (20, 20, 1),
max_neighbours = 12, cov_model = cov1)
write_property(prop_result, "R_OK_with.INC", "OK_RESULT", -99)


*Thanks and Regards,
sonal.