Zoning of the 10-Days, Monthly and Annual Rainfall Using GIS: A Case...

Zoning of the 10-Days, Monthly and Annual Rainfall Using GIS: A Case Study on Golestan Province, Iran

SHARE

Dr.HOSSEIN SHARIFAN
Faculty member of Water Engineering Department of Gorgan University, Iran

SINA ALAGHMAND (Corresponding Author)
M.Sc. Student of Civil-River Engineering
USM Engineering Campus, Nibong Tebal, 14300, Pulau Pinang, Malaysia
[email protected]

Abstract
Because of variation of climate conditions, from humid to arid, raining plays an important role in Golestan province. The South band of Golestan province is covered by density forests and croplands’, so in this case raining is defined as an important potential water resource for vegetations demands. On the other hand the North band of Golestan province with semi-arid and arid climate condition, raining is significant for water supply. In this research the variation of raining patterns in above mentioned province was studied. In order to create 10-days raining, initially the values of 10-days and monthly raining for 35 meteorological stations with 31 years period data were collected. To estimate 10-days and monthly raining values, the Minitab-13 software was used. Then the monthly predicted values in previous step were converted to 10-days raining according to distribution pattern. Therefore, the 10-days predicted raining values were calculated for each meteorological station for 2001 and 2002. By using Ilwis3.1, the results of previous step were visualized in GIS format. In this study for 10-days, monthly and annual raining, 72, 24 and 2 maps were created, respectively. According to this research, these can be concluded that firstly the amount of raining is decreased from South to North and also the distribution of raining becomes worse and secondly the prediction error for annual raining is less than monthly and also monthly is less than 10-days raining

Introduction:
Golestan province with 20380 km2 area is one of the most important provinces from agricultural aspect in Iran. Because of variation of climate conditions, from humid to arid, raining plays an important role in Golestan province. The south band of Golestan province is covered by density forests and croplands’, so in this case raining is defined as an important potential water resource for vegetations demands. On the other hand the North band of Golestan province with semi-arid and arid climate condition, raining is significant for water supply. Also for some objectives such as optimization of cultivation pattern in large croplands in Gloestan province, the predicted values of raining are significant. In this research Minitab-13 and time series were used to estimate the raining in studied meteorological stations. The aim of using time series was assigning of pattern and recognition of its behavior for prediction. The variables in a random process may be independent or dependent on each other. One of the primary factors for using data in time series process is the stability of data; otherwise the instability should be solved. In order to this aim the differential method and Box-Cox method can be used to stabilize the mean and variation, respectively. In order to create the time series models such as MA, ARMA, ARIMA, SARIMA and etc, there are 4 steps as follows: assigning the model, training the model, testing the model and predicting. Recently, ARMA, ARIMA and SARIMA are used for modeling and simulating of climatologic parameters. Some evidences are as follows: Jones (1986), Honson & Lebdof (1988), Blophil & Nitchka (1992) and Foland (1990). In Iran, because of the short precedence of using direct measuring of meteorological data, these techniques has restricted uses, here we can mention some related studies in Iran: Jamshidi (Tehran temp-rain model) and Maleki (West of Iran temp-rain model). Also Kheradmandnia and Asakereh used SARIMA model for prediction of monthly mean temperature in Jaask area.

Ashgar Toosi, by using time series predicted the drought in Khorasan province and according to the results suggested the best cultivation patter. Also Ahmadi used SARIMA model for prediction of annual rain of meteorological stations in Khorasan province. To select the best prediction method, it is necessary to consider many factors such as the objectives of the project, accuracy and budget. For example Neilor and colleagues proved that Box-Jenkinz one-variable prediction model is better than Whorton economical rate pattern. Box-Jenkinz method is based on a mental principal who allows us to use wide range of patterns.

On the other hand, for zoning of raining values it is possible to use Geography Information System (GIS). GIS is defined as a collection which provides some facilities such as accumulation, storage, management, modifying, analysis and visualizing of data for users. GIS was utilized in irrigation and related fields, as an evidence we can notice the Skop, E. and Acqaron, M. (1997), who used GIS for mapping of evapo-transpiration in the Vejle Fjord watershed, Denmark. They applied the soil suitability map, time series of monthly rain and potential evapo-transpiration in a regression equation and then estimated the spatial distribution of real evapo-transpiration. In the above mentioned study, GIS was used for interpolation of monthly raining between index meteorological stations and to combine them with soil suitability map, which lead to real evapo-transpiration map.

Another related study was done in Okatagan valley in USA for planning of orchards and croplands. Also Rojaz and Roldan by using evapo-transpiration map in Jean province of Spain, planed for irrigation of olive orchards. In Iran number of researchers such as Asadi and colleagues used GIS for zoning of irrigation data in Khoozestan province.

As it clears the rate of rain is one the most important factors for optimization of cultivation pattern. So one of the main objective of this research was prediction of 10-days, monthly and annual raining in meteorological stations. On the other hand, zoning of raining in 10-days, monthly and annual using GIS was the other objective of present paper.

Material and methods:
In statistical researches the existence of correct and long period statistics and data, is a significant parameter. Therefore according to the objectives of this research, attempted to use maximum number of meteorological stations with reliable and long period data, in this case 35 meteorological stations with 31 similar year’s data were collected. In order to predict raining, 10-days and monthly data were classified in columns in a Q-Basic program. Then by means of Minitab-13 and SPSS, the raining values for the next 2 years were created.

In order to calibrate and check the correctness of the model and the predicted raining values, the 10-days real values of the 4 last years were compared with the predicted ones and the best model was selected. So, first 10-days and monthly values raining were predicted according to the recorded data, and then monthly predicted values were distributed to 10-days raining. Therefore in order to create the 10-days raining in 35 meteorological stations in Golestan province, the above mentioned method was used, so 4 steps were applied, so that for recognition of variance stability of main raining data the Bartlet and Loan tests (in Minitab-13) were used and in case of instability the Box-Cox conversion was used.

In order to select the final pattern and probable corrections in primary pattern, situation of remained values of considerable patterns, number of parameters in each pattern and correlation between parameters were examined. In last step, on basis of final model for all meteorological stations the 10-days raining values were created and the annual raining for next 2 years (2005 and 2006) were estimated according to the accumulation of monthly raining values. After these steps, zoning of 10-days, monthly and annual raining were done in GIS.

For visualization the results of this research and creating the attribute maps, 3 GIS software’s were used, Arcview3.1, Ilwis3.1 and R2V. In this case, first by means of Ilwis3.1 the latitudes and longitudes of meteorological stations were converted to UTM projection (X, Y), and saved in data base files (.dbf format). In the next steps the X, Y coordinate of meteorological stations were imported (Add) in Arcview3.1. In Arcview3.1 point layer of each time series were created and then in order to create the raster layers, the point layers imported to Ilwis3.1 with .shp format. Finally in Ilwis3.1 by using Moving Average method from Interpolation menu, the attribute raster map for each time series was created.

Conclusion and Results:
Table 1 is showing the results of this research for Gorgan synoptic station for 2006 in mm. After estimation of 10-days, monthly and annually raining values for all meteorological stations, all attribute maps of next 2 years (predicted values for 2005 and 2006) were created in Ilwis3.1. It means that 2, 24 and 72 maps were created for 10-days, monthly and annually raining, respectively. As an example figures 1 and 2 are illustrating the zoning raining maps for third decade of September of 2005 and annually raining of 2005 in Golestan province.

Table 1: Results for Gorgan Synoptic Station.

According to this research these can be concluded that firstly the amount of raining from South to North is decreased and also the distribution of raining becomes worse and secondly the prediction error for annual raining is less than monthly and also monthly is less than 10-days raining.


Figure 1: Zoning raining maps for third decade of September of 2005 in Golestan province in Ilwis3.1


Figure 2: Zoning raining maps for annual raining of 2005 in Golestan province in Ilwis3.1

References:

  1. Abu Zreig, M., M. Attom, and N. Hamasha. 2000. Rainfall Harvesting Using Sand Ditches in Jordan. Agricultural Water Management 46:183-192.
  2. Arabkhedri, M., A. Sarreshtehdari, and K. Kamali. 1997. The Long Time Effect of Flood Harvesting on Infiltration Rate, p. 1157-1158 Proceeding of the 8th International Conference on Rainwater Catchments System, Vol. 2. Ministry of Jihad-e-Keshavarzi, Tehran, I.R. Iran.
  3. ILWIS 3.2. 2001. ILWIS 3.2 Academic User’s Guide ITC, Enschede.
  4. Study group. 1995. Flood Spreading project in Abbarik-Bam-Kerman (Persian). Soil Conservation and Watershed Management research Center.
  5. Sharma, K.D., H.P. Singh, and O.P. Pareek. 1983. Rainwater infiltration into bare loamy sand. Hydrological Sciences Journal 28:417-424.
  6. Shahoie, S & Refahi, H. 1996. Using Meteorological Data in Estimation of Erosion, 1st Seminar on Erosion and Sediment, Noor, Pages: 265-273.
  7. No name. Atlas of Geography Rivers of Iran, 2nd volume. Caspian Sea River Basin. Press of Defense Ministry of Iran.
  8. Yousefi, A, R. 2003, Phiziography of Golestan Province River Basins.