Mann Kendall Test

AhsanAbbas

Hello Guys, I need to perform Mann Kendall test on 5 series raster data yearly basis, from 2007 to 2011.

if need to iterate pixel by pixel on raster cell values for calculating the sign. How can i do it... Below is code but i have problem with numpy...

import arcpy
import numpy
from arcpy import env


env.workspace = r"D:\PythonProject\Data"
rasterlist = arcpy.ListRasters()


for raster in rasterlist:
    print raster


n = len(rasterlist)
pairs = (n * ( n - 1)) / 2
print "Pairs to be formed: ", pairs


def pairsRasters(raster1, raster2):
    print raster1, " - ", raster2
    
    inRas1 = arcpy.Raster(raster1)
    lowerLeft1 = arcpy.Point(inRas1.extent.XMin,inRas1.extent.YMin)
    #cellSize = inRas.meanCellWidth
    my_array1 = arcpy.RasterToNumPyArray(inRas1,nodata_to_value=0)


    inRas2 = arcpy.Raster(raster2)
    lowerLeft2 = arcpy.Point(inRas2.extent.XMin,inRas2.extent.YMin)
    #cellSize = inRas.meanCellWidth
    my_array2 = arcpy.RasterToNumPyArray(inRas2,nodata_to_value=0)
    


for current in reversed(rasterlist):
    for previous in rasterlist:
        if current == previous:
            pass
        elif current > previous:
            pairsRasters(current, previous)
        else:
            pass

Suppose the two raster of similar extent, i need to calculate difference, if it's greater than 1 the value store in output raster would be 1, if less than 1 the value would be -1 and else zero. How can i iterate on pixel by pixel of two rasters. so i can save it into new raster. That's it..

Column 1	Column 2	Column 3	Column 4	Column 5
50	25	98	102	98
122	98	10	210	100
57	66	92	21	10
98	45	255	9	59
58	98	17	31	54

Message was edited by: Ahsan Abbas

newData.zip

DanPatterson_Retired

try the following to see if this is what you want

3 arrays with differing values and sign

>>> a
array([ 6.,  5.,  5.,  3.,  1.,  1.])
>>> b
array([ 1., -2.,  3.,  4.,  5., -6.])
>>> c
array([-3., -2., -1.,  4., -5., -6.])
>>> np.sign(b-a)
array([-1., -1., -1.,  1.,  1., -1.])
>>> np.sign(c-b)
array([-1.,  0., -1.,  0., -1.,  0.])
>>> np.sign(a-c)
array([ 1.,  1.,  1., -1.,  1.,  1.])
>>>