120 
SOIL SALINITY ESTIMATION IN SOUTHERN ALBERTA USING RADAR DATA 
C.A. Hutton 
INTERA KENTING 
1525 Carling Ave. 
Ottawa, Ontario 
R.J. Brown 
Canada Centre for Remote Sensing 
1547 Merivale Rd. 
Ottawa, Ontario 
M.G. Lowings 
Polar Sea Research 
204, 740 - 4 Avenue South 
Lethbridge, Alberta 
Soil salinity is an increasing problem in the 
Canadian prairies. Irrigation, in addition to many 
popular farming practices such as summerfallowing 
have increased and intensified the problem. Since 
land reclamation is difficult and expensive remote 
sensing offers a potential tool with which to monitor 
the problem and allow more accurate and timely 
remedial measures. Since areas of salinity will move, 
shift, be introduced or increase in severity, as 
changes occur in the rainfall and groundwater 
patterns, an adequate monitoring method becomes 
critical. 
Previous efforts to identify saline soils with VIR 
data have indicated some intermediate to mature 
saline seeps can be detected by salt crusting or 
identification of indicator plant species. 
(Chaturvedi, L., et al., 1983; Dalsted, K.J., 1979) 
Since, SAR offers a unique interaction with ground 
targets, the potential for identification of saline 
areas and their severity is unknown. SAR data was 
acquired over an area surrounding the town of 
Warner, in southern Alberta to assess its potential 
for mapping and monitoring salinity. The objectives 
of the present work were to assess 1) the capability 
of SAR data for saline area identification and 2) 
determine the potential for mapping the severity of 
these areas. 
C and X-Band narrow mode HH data was acquired in 
July while both HH and VV were collected in October 
of 1989. The July data were analyzed to locate 
saline areas through varied backscatter related to 
salt crusts (or chemical variations in the soil) or 
differing dielectric constants which might indicate 
less severe areas. In vegetated areas a reduction 
in crop health, identified through reduced 
backcscatter, hopefully will correlate with salinity. 
Using the October data set, collected after harvest, 
the indicators of salt crusting were assessed in 
addition to the location of two saline indicator plant 
species Kochia and Foxtail. Due to the absence of 
other vegetative growth (crops) these plant species 
were felt to be a means of determining saline areas. 
The methodology of combining the poor growth areas 
on the July imagery with regions with indicator 
species was developed as a means of saline 
identification. The C and X-Band data were used 
both to compare the information found with each 
frequencies and determine a preferred frequency or 
combined frequency for mapping. Also, HH and VV 
were both assessed addressing any polarization 
dependence that might occur. 
A number of preprocessing tasks were carried out 
in preparation of this data analysis. The data were 
radiometrically corrected to remove the antenna 
pattern effects and were filtered with a 5x5 median 
filter to remove the speckle. The data were 
stretched to fill the entire dynamic range or 
enhanced to provide the greatest visual contrast 
between the features on these images. Negatives 
were created from these data and reproduced 
photographically at an enlarged scale for use in the 
analysis. The analysis was conducted qualitatively 
in the first stages using photo interpretation but 
will be expanded to include digital methods at a 
later data. 
A SUN system running the PCI EASI/PACE software, 
was used to carry out the preprocessing. A 
Compaq-base EASI/PACE package and a PAMAP GIS 
package was used to conduct the mapping portions 
of this experiment. A map was produced identifying 
the crop types and medium and severe saline areas. 
The salinity information was digitized from a Soil 
survey of the County Warner, Alberta, 1984. 
Using the saline map and field overlay, the ability 
to identify these known saline areas were 
determined. The factors influencing the radar 
backscatter such as crop type, topography and 
hopefully salinity were assessed. The regions of 
known salinization were assessed to determine 
whether the SAR could identify these and map any 
changes which have occurred. 
Outputs from the analysis will include maps 
indicating saline areas derived from the SAR 
imagery and, SAR/VIR imagery in combination. The 
mapping potential of the SAR data set in reference 
to frequency, polarization, incidence angle and date 
are discussed. 
REFERENCES 
Chaturvedi, L., K.R. Carver, J.C. Harlan, G.D. 
Hancock, F.V. Small and K.J. Dalstead, 1983, 
"Multispectral Remote Sensing of Saline Seeps", IEEE 
Transactions on Geoscience and Remote Sensing, Vol. 
GE-21. No. 3, pp 239-250.