Research Article
Past Runoff Trend for Declivitous Farms of Iran
Department of Agronomy and Plant Breeding, Shahrood University of Technology, P.O. Box 36155-316, Shahrood, Iran
There is a general consensus that global average surface air temperature has increased during the 20th century (Houghton et al., 2001). In most cases, the value of precipitation has changed dramatically. For example, Dai et al. (1997) found that on a globally averaged basis, precipitation over land has increased by about 2% over the years 1900-1997. Zonally averaged precipitation has increased by 7-12%, between 30 °N and 85 °N, compared with a 2% increase for 0-55 °S (Folland et al., 2001). Groisman et al. (2004) reported increases in precipitation over the conterminous USA during the 20th century, with most of the increase confined to the spring, summer and fall. Brown (2000) found systematic increases in winter (December through February) precipitation over North America averaging 3.9% per decade over the period 1915 through 1992. Mekis and Hogg (1999) showed that precipitation in Canada has increased by an average of more than 10% over the 20th century. Zhai et al. (1999a) reported the area affected by the upper 10% of heaviest precipitation has significantly increased. It has also been found an upwardly change in precipitation for many places, including west China (Zhai et al., 1999b), northern Europe (Schonwiese and Rapp, 1997), former USSR (Groisman and Rankova, 2001), Australia (Hennessy et al., 1999) and Southern Hemisphere (Dai et al., 1997).
On the other hand, some reports indicated the declining trend in precipitation during past decades. For example, an analysis of rainfall data since 1910 by Haylock and Nicholls (2000) reveals a large decrease in total precipitation and related rain days in southwestern Australia. Over the last 50 years, there has been a slight decrease in annual precipitation over China, which is supported by a significant (5% confidence level) decrease in the number of rainy days (3.9% per decade) (Zhai et al., 1999a). There have also been marked decreases in precipitation in some other regions, like southern Europe (Schonwiese and Rapp, 1997), Southeast Asia and the western and central South Pacific (Manton et al., 2001) and East Africa (Mwale et al., 2004).
It has been found by Georgievsky et al. (1996) the increases in precipitation over the last several decades over western Russia, accompanied by increases in stream flow (runoff) and a rise in the level of the Caspian Sea. Multidecadal stream flow data in Canada have revealed that there are no apparent inconsistencies between observed changes in stream flow and precipitation (Zhang et al., 2000). Published reports indicating that runoff from the Mississippi river increased by 22% from 1949 to 1997, during which time precipitation increased by 10% (Milly and Dunne, 2001). Several analyses (Groisman et al., 2001) have detected increases in stream flow across much of the contiguous United States, confirming the general tendency to increasing precipitation. Stream flow data for major rivers in southeastern South America for the period 1901 to 1995 show that stream flow has increased since the mid-1960s, which is in consistent with the changes in precipitation (Garcia and Vargas, 1998). Increases in precipitation over land have been associated with corresponding increases in runoff in river basins in the conterminous USA (Groisman et al., 2001). In Canada, increasing temperature combined with almost no change in precipitation, resulted in no change in annual stream flow from 1947 to 1996 for most regions (Zhang et al., 2001). In contrary with these findings, some reports indicating the inconsistencies between rainfall and runoff. As instance, Lins and Michaels (1994) found in some regions that changed stream flow did not relate well to an increase in rainfall; in Tien Shan Mountains in Northern Eurasia, precipitation has increased, snowfall decreased and runoff decreased or not changed during the latter half of the 20th century (Aizen et al., 1997). The reasons for inconsistencies between observed changes in stream flow and precipitation can be found in Huntington (2006).
The published reports regarding trends in runoff are rare for Iran. In one location of Iran, i.e., Gorgan, Gholipoor and Soltani (2005) has studied the changes in monthly value of runoff across years 1961-2001; in the procedure of calculation of runoff which they have used, the effect of some factors on runoff, including evaporation, snow-falling and snow-melting, has not been included. Therefore, the objective of this study was to evaluate the trends in runoff, using more efficient procedure; this evaluation was supposed to be in monthly and yearly point of view and in view point of growing period of chickpea for five locations of Iran. The result of this study may be useful for updating the recommendable value of slope for planting the crops, including chickpea, in declivitous farms.
Model Description
A simulation study was conducted to determine the past changes in runoff for Iran in 2006-2007 at Shahrood University of Technology. In this study, the soil water balance sub-model of CYRUS model was used to calculate the daily value of runoff; firstly, it was recoded in Qbasic programming language and run for medium deep silty loam soil. This model was initially designed in 1999 by Soltani et al. (1999) for chickpea (Cicer arietinum). Then it was developed for seedling emergence (Soltani et al., 2006d), for leaf expansion and senescence (Soltani et al., 2006c), for response of leaf expansion and transpiration to soil water deficit (Soltani et al., 2000), for response to photoperiod (Soltani et al., 2004), for harvest index (Soltani et al., 2005), for phenological development (Soltani et al., 2006a), for nitrogen accumulation and partitioning (Soltani et al., 2006b) and for the effect of temperature and CO2 (Soltani et al., 2007).
The CYRUS has been used for evaluating yield of chickpea and its stability in dormant seeding (Soltani and Torabi, 2007), determining optimum phenology of chickpea for now and future (Rahimi-Karizaki and Soltani, 2007), study of past climate change effects on chickpea phenology at different sowing dates (Gholipoor and Shahsavani, 2008), potential effects of individual versus simultaneous climate change factors on growth and water use in chickpea (Gholipoor, 2007), evaluating the effect of future climate change on yield of rainfed chickpea in northwest of Iran (Barzegar and Soltani, 2007), comparing relative effects of temperature and photoperiod on development rate of chickpea (Gholipoor and Soltani, 2006), optimizing the dormant sowing of chickpea (Gholipoor et al., 2006) and finding future climate impacts on chickpea in Iran and ICARDA (Gholipoor and Soltani, 2008). The soil water balance sub-model of this model with some modifications has been applied for comparative evaluating the climate-related runoff production in slopped farms of Iran (Gholipoor, 2008a).
Briefly, in soil water balance sub-model, daily soil water content is estimated as fraction transpirable soil water [FTSW, which ranges from 0 (point at which plants face with wilting) to 1 (field capacity)]. Similar to that described by Amir and Sinclair (1991), it accounted for additions from infiltration and losses from soil evaporation, transpiration and drainage. Infiltration is calculated from daily rainfall less any runoff. Runoff is estimated using the curve number technique (Knisel, 1980). Soil evaporation (E) is calculated using the two-stage model as implemented in spring wheat model developed by Amir and Sinclair (1991). Stage I E occurs when water present in the top 200 mm of soil and FTSW for the total profile is greater than 0.5. Stage II E occurs when the water in the top layer is exhausted or the FTSW for the total soil profile reaches to less than 0.5. In stage II, E is decreased substantially as a function of the square root of time since the start of stage II. The calculation of E is returned to stage I only when rain or irrigation of greater than 10 mm occurs. Like procedure of Tanner and Sinclair (1983) and Sinclair (1994), the daily transpiration rate is calculated directly from the daily rate of biomass production, transpiration efficiency coefficient (5 Pa) and VPD. The calculation of VPD is based on suggestion of Tanner and Sinclair (1983) that it to be approximately 0.75 of the difference between saturated vapor pressure calculated from daily maximum and minimum temperatures. Briefly description for other sub-models can be found by Gholipoor and Soltani (2008).
In this investigation, as mentioned previously, the sub-models regarding growth and development of chickpea were inactivated and only soil water balance sub-model was used; therefore, like previous work (Gholipoor, 2008a), some modifications were made as follows: (1) transpiration was assumed equal to zero, (2) a sub-model was added for calculating snow cover and snow melting as report of Ritchie (1991); this procedure has been used by many researchers including Soltani et al. (2006d); in this sub-model the value of snow is calculated based on maximum temperature; this temperature is also used for daily calculating amount of snow melting and (3) in original model, time course for calculations is growing period of chickpea; here it was also considered for monthly and yearly periods. It should be mentioned that the planting date for chickpea was considered as the 7 days with no rainfall and with mean temperature above the base temperature; then growing period was calculated on the basis of required growing degree days from planting to maturity.
Locations, Measured Attributes and Determining the Trends
Five locations with long-term and reliable daily weather data were selected for the study to represent a large geographical area and several climatic zones in Iran (Fig. 1 in Gholipoor (2008a)). The selected sites included Isfahan (32.67 °N, 51.87 °E and 1600 m asl), Shiraz (29.55 °N, 52.60 °E and 1488 m asl), Kermanshah (34.32 °N, 47.12 °E and 1322 m asl), Tabriz (38.13 °N, 46.28 °E and 1364 m asl) and Mashhad (36.27 °N, 59.63 °E and 990 m asl). Mean annual temperature is 16.4°C for Isfahan, 17.8°C for Shiraz, 14.4°C for Kermanshah, 12.6°C for Tabriz and 14.3°C for Mashhad. The annual rainfall occurs during 35 wet days for Isfahan, 42 for Shiraz, 74 for Kermanshah, 81 for Tabriz and 63 for Mashhad. For each location, 39 (1966-2004 for Tabriz) to 44 years (1961-2004) of daily data for rainfall and maximum and minimum temperatures and sunshine hours were available. Solar radiation data were calculated from sunshine hours and extraterrestrial solar radiation as outlined by Doorenbos and Pruitt (1977).
The calculated main attributes were runoff, number of days with runoff (R), frequency of R greater than zero and lower than and/or equal to 2 mm (0 < R ≤ 2), of 2 < R ≤ 4, of 4 < R ≤ 6 and of 6 < R ≤ 8. Some other attributes were also calculated, including mean temperature, solar radiation, evaporation, probability of occurrence of rainfall (Ra), amount of Ra, number of days with ≤ Ra (wet days), frequency of Ra≤5 mm, of 5 < Ra→10, of 10 Ra ≤15, of 15 Ra ≤ 20 and of Ra>20. Determination of probability of occurrence was achieved using generation of cumulative distribution functions (Purcell et al., 2003). Calculation of main attributes and of evaporation was done for sloped farms. The value of slope was considered to be 14%; because cultivation of crops is usually not recommendable in slops steeper than 14%.
If the changing trend of runoff, for example, to be increasing across first 20 years, but decreasing across last 24 years, it may be described by quadratic function; such changes may be considered as periodic-fluctuation, rather than consistent change, in runoff which is described by linear function. Therefore, as it has been done for temperature by Gholipoor (2008b) and for some climatic variables by Soltani and Soltani (2008), it was only used a simple linear regression function (Y = a+bX) to determine trends (slopes), i.e., b, in runoff and in other attributes; in this equation, Y is dependent variable (like runoff), X year, a intercept and b slope of regression line; the value of parameters (a and b) was calculated, using the procedure REG in SAS (1989). Like many researchers, including Soltani and Gholipoor (2006), the rates with probability level equal to and/or lower than 0.1 were considered as significant.
The averaged values of runoff for slope 14% were presented in Fig. 1. In Isfahan, the monthly value of runoff was relatively considerable (1.03 to 2.29 mm) for months January to April; while it tended to be negligible (0.2 to 0.3 mm) for May, November and December; it was zero for the rest months of the year. Among locations, the highest monthly value of runoff (32 mm) was found for Shiraz (in January); in this location, amount of runoff declines to zero over May, June, August, September and October. In Kermanshah, that of runoff ranged from 10 to 15 mm over first four months of year and from 1.5 to 7.3 over May, October, November and December. Runoff tended to vary between 3.2 to 5.3 mm over March, April and May in Mashhad; it was little in firs two months of year and negligible in last two ones. In Tabriz, despite of above named locations, the average value of runoff does not decline to zero over summer and other months of the year; the probability of raining over July, August and September is 8.03% for this location, which is considerably higher than 0.54 to 2.83% for other locations; in Tabriz, like Kermanshah, the pick value of runoff is seen in March. The detailed cause of these mentioned variations between locations for value of runoff, including ratio of snow to rainfall, evaporation and occurrence of shallow, medium and heavy rainfalls, can be seen in report of Gholipoor (2008a).
The rate of change (value of b in equation Y = a+bX) and its probability level for 6 variables, including value of runoff (R), number of days with R, frequency of R greater than zero and lower than and/or equal to 2 mm (0 < R≤2), of 2 < R≤4, of 4 < R≤6 and of 6 < R≤8 for Isfahan was shown in Table 1. In January, consistent with increased rainfall (Table 2), the past changes in value of runoff have been upwardly (2.64 mm per 44 years); frequency of 4 < R≤6 (almost heavy runoffs) found to be decreasing associated with corresponding downwardly trend for frequency of Ra>20 (Table 2); based on this change and on increasing trend of number of days with runoff, it can be cleared that incline in runoff has been due to enhancement in frequency, rather than intensity, of runoff. In February, the changes in rainfall and consequently in runoff were not sensible; just it was found statistically significant decrease in frequency of 0 < R≤2. The ratio regarding rate of increase in rainfall to that of incline in runoff tended to be 8.33 in March, which is considerably higher than that in January (4.76); in other words, the rate of increase in precipitation was higher in March as compared to that in January; while, that of incline in runoff was unexpectedly lower in March, when compared with that in January; part of this can be attributed to evaporation which its upwardly trend appeared to be statistically considerable in March, but negligible in January (Table 2); it should be mentioned that the change in evaporation for March has been due to increase in solar radiation, not to that in temperature (Table 2). In April, little (statistically negligible) diminishing trend in precipitation and on the other hand, little (negligible) increasing trend in evaporation have caused that decline in runoff to be huge (3.04 mm per 44 years); this decrease has been accompanied by decline in frequency of 6 < R≤8 (0.26 day per 44 years) and in number of days with runoff (1.28 days per 44 years). As mentioned previously, over June, July, August, September and October, the rare rainfall (the probability of occurring is about 2%) tends to infiltrate completely, even into the soil of declivitous (14%) lands. The value of runoff has statistically been constant for November and December; these months have never experienced runoff greater than 2 and lower than and/or equal to 8 mm. Due to increased evaporation by 0.502 mm year-1, runoff has been decreasing for growing period of chickpea. In yearly point of view, the value of runoff has changed from 5.29 mm in 1961 to 9.25 mm in 2004 (about 0.09 mm year-1), which is in agreement with reports for other countries (e.g., Groisman et al., 2001; Zhang et al., 2000).
Fig. 1: | The monthly sum value of runoff for slope 14% in five locations of Iran |
Table 1: | The rate of change and its probability level (P) for value of runoff (mm year-1) and for number of days with runoff (R>0) (day year-1), for frequency of runoff (R) greater than zero and lower than and/or equal to 2 mm (0 < R≤2; day year-1), of 2 < R≤4, of 4 < R≤6, of 6 < R≤8 across 1961-2004 in Isfahan, Iran |
#: Growing period of chickpea (day of year 58-176). The bold values are significant |
Table 2: | The rate of change and its probability level (P) for value of evaporation (mm year-1) and of rainfall (mm year-1), for number of days with rainfall (wet days; day year-1) and with heavy rainfalls (rainfall>20 mm; day year-1), for mean temperature (°C year-1) and for solar radiation (MJ/m2/day/year) across 1961-2004 in Isfahan, Iran |
#: Growing period of chickpea (day of year 58-176). The bold values are significant |
In Shiraz, the probable stimulating effect of increased frequency of 0 < Ra≤5 (data for frequency of different values of rainfall, for wet days, evaporation, rainfall, mean temperature and solar radiation not shown for Shiraz and for other locations which are discussed latter) on runoff might be counterbalanced by enhanced evaporation for January; therefore, the value of runoff in this month, which is the highest, has fortunately not increased (Table 3). In February, there have been no change in rainfall and hence in runoff; this was true for evaporation, but not for solar radiation and mean temperature; they have had directed change with incrementing the year. In March, the increase in frequency of 10 < Ra≤15 has been accompanied by incline in that of 0 < R≤2 during 1961 to 2004; despite of this change, amount of runoff has been the same; this was also true for evaporation. In April, just frequency of 6
Table 3: | The rate of change and its probability level (P) for value of runoff (mm year-1) and for number of days with runoff (R>0) (day year-1), for frequency of runoff (R) greater than zero and lower than and/or equal to 2 mm (0 < R →2; day year-1), of 2 < R →4, of 4 < R → 6, of 6 < R → 8 across 1961-2004 in Shiraz, Iran |
#: Growing period of chickpea (day of year 54-172). The bold values are significant |
Table 4: | The rate of change and its probability level (P) for value of runoff (mm year-1) and for number of days with runoff (R>0) (day year-1), for frequency of runoff (R) greater than zero and lower than and/or equal to 2 mm (0 < R → 2; day year-1), of 2 < R → 4, of 4 < R → 6, of 6 < R →8 across 1961-2004 in Kermanshah, Iran |
#: Growing period of chickpea (day of year 83-201). The bold values are significant |
The rate of changes in measured attributes regarding runoff for Kermanshah was presented in Table 4. The value of b for runoff and other attributes was statistically zero in January and March. That of runoff, which is relatively considerable (averagely 9.96 mm) in February, has fortunately diminished; this has just been due to the optimized distribution of rainfall (decreased frequency of Ra>20 by 0.06 day year-1, with no change in value of rainfall). In April, the considerable decrease in rainfall has been accompanied by remarkable decline in frequency of 0 < R →4; despite of this change, amount of runoff has statistically been remained constant. The month May tended to be characterized as the month with no happening of 6 < R →8; the changes in runoff and frequency of 0< R≤6 have been diminishing for this month; part of the reasons for these changes may be due to diminished value of rainfall (0.71 mm year-1); warming the temperature (about 0.03 °C year-1) and consequently intensified evaporation (0.42 mm year-1) has also been cause of the named changes. In October and November, despite of warming the temperature by 0.06 and 0.04°C year-1, respectively, the value of evaporation, rainfall and hence, that of runoff appeared to show constancy. The month December was characterized by worsening of rain distribution, i.e., increased frequency of Ra>20, with no change in value of rainfall; this situation has caused that the frequency of 4 < R →6 and as a result, the value of runoff to have directed change with incrementing the years from 1961 to 2004. The downwardly change in frequency of 0 < R → 4 and in amount of runoff found to be significant for growing period of chickpea, due to diminished rainfall. The results regarding yearly calculations indicated that among attributes presented in Table 4, the decreasing trend has statistically been true just for frequency of 0 < R < 2 and 6 < R → 8.
The value of b for measured attributes regarding Tabriz was shown in Table 5. For January, the decrease in value of rainfall by about 0.31 mm year-1 with no change in frequency of Ra>20, have resulted in that frequency of 4 < R →6 to show downwardly trend; despite of this lessened frequency of almost heavy runoffs, amount of runoff has been the same across years 1966 to 2004; that of evaporation found to be not affected by significant warming and intensified reaching of solar radiation to soil surface (0.05 MJ/m2/day/ year) during named years. As compared to January, the rising slope of mean temperature and solar radiation has been more considerable in February; therefore the increase in evaporation found to be considerable (0.54 mm year-1) in this month, but as mentioned, negligible in January; declined rainfall and such change in evaporation for February have caused that the risk of runoff to be alleviated (i.e., decreased runoff). In March, June, August, September, November, December and yearly period, the value of b was statistically zero for value of runoff and for frequency of different values of runoff; in most cases, this might be due to no change in rainfall (September and yearly period with declined rainfall were exceptions). Because of significant changes in rainfall intensity, the frequency of 4 < R →6, of 2 < R→4 and of 0 < R →2 have being rising for months April, May and July, respectively; the amount of runoff has been the same for named months, due to no change in amount of rainfall. In growing period of chickpea, it was found a steady state for value of runoff and for frequency of calculated values of runoff.
As it was presented for Mashhad in Table 6, the value of runoff has had a plateau state across the years 1961 to 2004 for January; this has been associated with corresponding plateau state for rainfall and evaporation; on the other hand, the intensity of runoff has been changed as decrease in frequency of 4 < R →6, but rise in that of 0 < R →4; this has been accompanied by decrease in frequency of Ra>20. The increase in solar radiation and consequently intensified evaporation appeared to be significant for February and March; despite of these changes, the amount of rainfall and hence that of runoff have remained constant; positively correlated with frequency of different values of rainfall, it found a constancy in frequency of different values of runoff. In April, the intensity of runoff has been changed as amplified frequency of 6 < R→8 and on the other hand, as deducted that of 0 < R →`2 and 4 < R→6; this has been related to decreased evaporation and to increased frequency of 15 < Ra →20; in contrary with these changes, the value of runoff has been constant. Corresponded with trivialized frequency of 15 < Ra →20, that of 4 < R→6 has been abated for May; on the other hand, happening of 2 < R →4 has been intensified. In November, the decreased amount of runoff may be attributable only to warming-resulted rises in loss of soil water through evaporation (0.19 mm year-1); such exclusive attributing is due to no changes in value and intensity of rainfall; the decreasing trend in runoff for this month is consistent with report of Aizen et al. (1997) for some regions of Tien Shan Mountains in Northern Eurasia. For December and yearly period, the value of b was statistically zero for all presented attributes in Table 6. In view point of chickpea growing period, frequency of 0 < R →2 and 4 < R →6 has been diminishing.
Table 5: | The rate of change and its probability level (P) for value of runoff (mm year-1) and for number of days with runoff (R>0) (day year-1), for frequency of runoff (R) greater than zero and lower than and/or equal to 2 mm (0 < R →2; day year-1), of 2 < R→ 4, of 4 < R→6, of 6 < R→8 across 1966-2004 in Tabriz, Iran |
#: Growing period of chickpea (day of year 95-207). The bold values are significant |
Table 6: | The rate of change and its probability level (P) for value of runoff (mm year-1) and for number of days with runoff (R>0) (day year-1), for frequency of runoff (R) greater than zero and lower than and/or equal to 2 mm (0 < R→2; day year-1), of 2 < R →4, of 4 < R→`6, of 6 < R → 8 across 1961-2004 in Mashhad, Iran |
#: Growing period of chickpea (day of year 71-189). The bold values are significant |
The results indicated that in Isfahan, the value of runoff has increased for January, March and yearly period; on the other hand, it has had a downwardly change for April; runoff has shown statistically steady state for the rest months of the year. It appeared to have upwardly trend just for December in Shiraz. In Kermanshah, it was found the increasing trend in runoff for December, but decreasing trend for February and May. The months February and November have experienced the increasing risk of runoff in Tabriz and Mashhad, respectively. Over growing period of chickpea, this risk has been diminishing for Isfahan and Kermanshah; while it found to be constant for other locations.
Plowing the soil in declivitous farms for planting the crops makes the soil more susceptible to runoff-resulted erosion. Therefore, the recommendable value of slope of land is determined using the value (risk) of runoff. Accordingly it can be concluded that the recommendable slopes for planting the chickpea in declivitous farms of Isfahan and Kermanshah is steeper in 2000s as compared to 1960s. Similar conclusion can be derived for other crops with different season of growth and development. In agronomy point of view, such changes may be useful to extend the planting area. While in more cases, the decrease in runoff is coincided with diminished rainfall; in such conditions, the strategies for decreasing the loss of soil water may be more important. Leaving the crop residue on soil surface could decreased evaporation; some other benefits can be defined as increase the soil infiltration and consequently decreased runoff, enhancing yield (Power et al., 1998), soil organic carbon (Clapp et al., 2000), soil nitrogen content (Kumar and Goh, 2000) and C/N ratio (Martens, 2000).