چكيده لاتين :
Introduction
Ground water quality influences multiple natural and human factors. To check how
changes place in the passing of time and in the context of a location can orient point
of view to this for use in future years. One of the important indicators to show the
quality of drinking water and agriculture is the amount of nitrate contained in it
through corruption and human and animal waste, industrial products and waste
water from agriculture entering the surface and subsurface water. Other elements
that lead to a reduction in the quality of water sources, is phosphates. Phosphate the
use agricultural fertilizers complementary on the crops land. The causes of leaching
water entered to surface water or entered into the groundwater. Improvements that
have taken place recently in the field of spatial statistics and also the replacement of
the areal variable, instead of a random variable, promote a variety of methods of
interpolation in GIS is causes. Interpolation, spatial continuous changes to the face of
a defined level incarnate. Various researchers have done, extensive research in the
field of application of the procedures-ground statistics on groundwater conditions
and the achieved to results diverse. In Spatial changes investigation in the
concentration of nitrate of alluvial aquifer the North of the river Tagus in Portugal,
used from Kriging method. The results of this study indicated the high concentration
of nitrate was in the West of studied region. In spite of all the progress the last few
decades, particularly with the development in the field of spatial modeling there is
space, browse the resources and various studies show that the use of any of the
models hit the ground statistics is depending on the characteristics of the studied
region and the examined parameters. According to the past research and reviews so
far specialized study not done in the field of comparison and evaluation of the Kriging
and Co-Kriging series models with respect to auxiliary parameters in the studied
region. This research opportunity in spite of the relatively reasonable rainfall, in
terms of the qualitative variables, has been associated with problems. Fixes them, the
causes do this research. Indiscriminate Use of chemical fertilizers and in higher
quantities than the rate required by the absorption of the product, in the special
areas of the studied region agriculture led to the accumulation of too much them in
the soil and washing them into the body of groundwater. This study is trying in spite
checking of change the amount of phosphate and nitrate while in the Gharehsoo
watershed, compare of different methods earth statistics and estimates of best
practices in the area, is also identified the most important parameters affecting the
quality of the water resources in the Gharehsoo watershed in Golestan province.
Materials and Methods
In this study, statistical information of groundwater qualitative parameters related to
Piezomertic well (2013 statistical years' regional water organization of Golestan
province), were used for the calculations. At first, all of the data in the Excel software
classification for several years, separate and then in the Arc-GIS, using wells
geographical coordinates, their position on the map characterizing and finally provide
the position map of Piezomertic wells. To check the statistical characteristics and
water quality data zoning used of the software Spss17, the GS + 5.3 and Arc GIS 10.
Kriging is advanced methods of interpolation and fitted for data that has been
defined of topical trends. Ordinary Kriging Typical considerable variable values in the
sampling points will assion weight to each of the samples to state the linear
combination of the varied same values in the surrounding parts. Estimation of the
sampling points in the simple Kriging form of a weight linear combination and in its
estimation process must be careful that the rated characteristics of the second rank
stagnation. On the method of universal Kriging interpolation on spatial structure of
regional variables must exist both algebraic and flexibility change components
concurrently. In this method, it is assumed that in addition to the spatial correlation
between the parts exist diversion, or process in the z values. In the absence of data
distribution is complex and difficult to fit them the usual used disjunctive Kriging
models. As in classical statistics, there is also multivariate methods on the Geostatistic
based on correlation between data can be used from Co-kriging method. Cokriging
method is four kinds of normal, simple, comprehensive and disjunctive. In this
method, one or more secondary variable that is associated with that variable, used
for interpolated. This method is also suitable for that area of the station is lacking, on
the premise that the correlation of variables between can be increased estimate
accuracy. In this study, to work the map of underground water resources quality as
· the dependent variable and the data of the digital land use, distance to road and
residential areas, elevation, the distance of the forest, the distance from the
farmland, correlation between under-cropping area and amount of fertilizer
consumption, population density, groundwater levels, soil and geology as well as the
effective parameters on the process of water quality. On the models used from the
best equation on the basis number describing in the model, i.e. the square of
corrected correlation coefficient (R2 adjusted) and amount of F statistic. Whatever in
equation, amount of corrected correlation coefficient is near-optimal to one equation
is more favorable. For statistical evaluation of the resulting model, due to the limited
number of data, was used from method internal cross evaluation. In this method,
deleted a data from all series and is done Multiple Linear Regression without
Interfering its data again.
Discussion
With the acceptable precision of (Kappa calculation coefficient 88%) Image
classification, was used satellite image of the year 2010 and also as well as maximum
likelihood method in remote sensing software environment in the preparation landuse
map. Mean and standard deviation of the samples in this study is respectively
against 13.27, 9.06 for nitrate and 0.27, 0.21 for phosphate. Before doing any
calculation, initially check out being normal data using Kalmogorov- Smiranov test. By
used logarithms, being normal data. After the optimizing data, in order to describe
the spatial continuity of variables is draws view variable half of data. The results show
that the spherical model in both the Kriging and Co-kriging methods for nitrate and
spherical model in the Kriging method and the exponential model in the Co-kriging
for phosphate are practice the best model. Validation results using any of the three
criteria in the case of nitrate indicated is that the Co-kriging method in comparison
with Kriging has less error. Co-kriging of universal, simple, ordinary, disjunctive and
Kriging of universal, simple, ordinary, disjunctive has respectively the lowest swing
and error. Since that universal Co-kriging method shows value lowest, as a proper
method of interpolation was used in computing spatial map of nitrate on Gharehsoo
watershed. Validation results using any of the three criteria in the case of phosphate,
indicating it is Co-kriging and kriging methods of the disjunctive, simple Co-kriging,
universal, simple, ordinary kriging, and simple, ordinary Co-kriging respectively has
the lowest amount of swing and an error. According to that method of disjunctive Cokriging
shows the lowest value as a proper method of interpolation was used in
computing the spatial map of phosphate on Gharehsoo watershed. With review of
sensitivity analysis results the using model explaining coefficient (R2) that after
every time of running the amount of R2 has been extracted and on the basis of the
amount of difference with the full data series the effect of independent variable is
calculated, the independent variables (correlation between the under-cropping area
density and the extent of cultivation of nitrogen fertilizer consumption) has the most
influence on soil nitrate, land-use and over pollution, respectively, a great impact on
the distribution of underground water nitrate in the study area. Trend review of
nitrate changes location from East to West and North to South the basin is emphasis
on nitrate concentrations high in the Gargan and northern regions shows the role of
human factors, industrial and urban development, harvesting too much of
underground water resources and the use of chemical fertilizers indiscriminate.
Multiple regression modeling results also confirms the same issue.
Conclusions
Results of multiple linear regression modeling indicate that the independent variables
of distance from the forest, geology and over population at the level of 99% and the
independent variable, the relationship between the under-cropping area density and
the amount of phosphate fertilizer consumption at the level of 95%, have significant
correlation with distribution of phosphate in the Gharehsoo watershed that results of
the sensitivity analysis of the model with the use of explaining coefficient also
confirms. The study region, in spite of the relatively reasonable ra infall, they have
problem water quality. So the quality management of this important can on the
shadow of feasibility and recognize steps from of planning priorities and operation
management is considered. Trend review of water quality conditions changes and
causes an issue it can greatly into the coherent vision of managers in planning are
extremely helpful. So offer that the similar research is done in the area with the use
of other water quality data and geo-statistic different ways for look at the valve of
planners, for use in future years.
