**18. Discussion**

Investigation of the results indicates that the elevation of resulted runoff from produced rainfall on the plots varies from 0.41 to 2.98 mm and runoff coefficient from 2.6 to %18.6 . Sediment concentration which is due to runoff and sediment production together is between 2.25 to 50.8 gr/lit and sediment production from 3.06 to 112.5 gr/m3. If unit change for a plot having the highest amount of sediments, the amount of sediment production for this type of rainfall is 1.125 ton/hectare which is considerable.

Kolmogrove-Smirnov Test shows that runoff and sediment production data have normal distribution. In independent t-test, with regard to the amount of significant for Levene Test which is lower than 0.05, the variance is not equal in two landuses, therefore with regard to significant of this part which is 0.028, there are differences between sediment production in two cultivations and rangeland landuses so that in dry-farming cultivation, the amount of sediment production is %50 more than rangeland which is due to decrease of soil structure stability due to yearly cultivation activities and overgrazing(Zhou et al., 2010). Of course deep ploughing decreases runoff volume but increases erosion and sediment production. Figure 12 shows sediment variations in two prominent landuses in the area.

**Variable Unit Landuse N Mean** 

Runoff observation time min

Moisture depth cm

Total runoff cm3

Runoff mm

Turbidity gr/lit

Plot sediment gr/m2

**18. Discussion** 

Table 7. Average runoff and sediment in dry-farming and range landuses

Figure 12 shows sediment variations in two prominent landuses in the area.

type of rainfall is 1.125 ton/hectare which is considerable.

Investigation of the results indicates that the elevation of resulted runoff from produced rainfall on the plots varies from 0.41 to 2.98 mm and runoff coefficient from 2.6 to %18.6 . Sediment concentration which is due to runoff and sediment production together is between 2.25 to 50.8 gr/lit and sediment production from 3.06 to 112.5 gr/m3. If unit change for a plot having the highest amount of sediments, the amount of sediment production for this

Kolmogrove-Smirnov Test shows that runoff and sediment production data have normal distribution. In independent t-test, with regard to the amount of significant for Levene Test which is lower than 0.05, the variance is not equal in two landuses, therefore with regard to significant of this part which is 0.028, there are differences between sediment production in two cultivations and rangeland landuses so that in dry-farming cultivation, the amount of sediment production is %50 more than rangeland which is due to decrease of soil structure stability due to yearly cultivation activities and overgrazing(Zhou et al., 2010). Of course deep ploughing decreases runoff volume but increases erosion and sediment production.

Runoff coefficient (%)

Dry-farming 36 8.42

Dry-farming 36 5.50

Range 33 6.36

Range 33 3.33

Range 33 1882.3

Range 33 1.88

Range 33 11.76

Range 33 12.22

Range 33 25.51

Dry-farming 36 1443.8

Dry-farming 36 1.44

Dry-farming 36 9.02

Dry-farming 36 22.37

Dry-farming 36 38.64

Fig. 12. Mean sediment concentration and sediment production in rangeland and dry-farming cultivation

In ANOVA Test, due to the fact that the level of significant is lower than 0.01, there are significant differences in runoff and sediment production between different slope classes. For determining which slope classes are different from each other, Clustering of Mean Method was used. The results of Duncan Method indicates the notable effect of slope on the amount of runoff and sediment production, so that each slope class is clustered in a separate group and is indicative of meaningful difference between slope classes. Figure 13 and 14 show mean amount of runoff and sediment in different slope classes, respectively.

Investigation of Effective Factors on Runoff Generation and

Sediment Yield of Loess Deposits Using Rainfall Simulator 145

In depended variables gr/lit gr/m2 Unit cm3

equivalent calcium carbonate % .28 (\*) .02 .06

Ca meq/lit .09 -.42 (\*\*) -.29 (\*) ++ Mg meq/lit -.23 (\*) -.22 -.23 ++ K meq/lit .29 (\*) .10 .16 + Na meq/lit .46 (\*\*) .18 .30 (\*) +

Sum of Cations meq/lit .36 (\*\*) .01 .14 CO3 meq/lit .(a) .(a) .(a) -- HCO3 meq/lit .20 -.01 .05 - SO4 meq/lit .32 (\*\*) .02 .11 -- Cl meq/lit .23 .20 .25 (\*) - Anion sum meq/lit .40 (\*\*) .11 .22 SAR .48 (\*\*) .27 (\*) .38 (\*\*) ESP .48 (\*\*) .28 (\*) .38 (\*\*) Illite % .12 .10 .13 Chlorite % .26 (\*) .13 .20 Kaolinite % -.53 (\*\*) -.20 -.33 (\*\*) Smectite % .17 .26 (\*) .29 (\*)

Mixed layer Clay Minerals % -.04 -.15 -.12 (\*\* ) Correlation is significant at the 0.01 level (2-tailed). (\*) Correlation is significant at the 0.05 level (2-tailed). Table 8. Pearson Correlation coefficient between different measured variables with runoff

volume, sediment production and sediment concentration

Total runoff Turbidity Plot sediment Depended variables

slope % .880(\*\*) .665(\*\*) .80 (\*\*) Land cover % -.298(\*) -.273(\*) -.36 (\*\*) A-horizon depth cm -.428(\*\*) -.107 -.25 (\*) humidity % .27 (\*) .24 (\*) .21 gr/cm -.11 .04 -.04 Bulk density 3 EC ms/cm .24 (\*) .04 .10 pH .30 (\*) .21 .31 (\*\*) Organic mater % -.31 (\*\*) -.16 -.28 (\*) CaSO4 meq/100s -.15 -.19 -.17 Sand % -.25 (\*) -.17 -.26 (\*) Clay % -.35 (\*\*) -.27 (\*) -.33 (\*\*) Silt % .54 (\*\*) .39 (\*\*) .53 (\*\*) CEC cmol/kg -.16 .46 (\*\*) .27 (\*)

Fig. 13. Mean runoff production in each slope classes

Fig. 14. Mean sediment production in each slope class

Fig. 13. Mean runoff production in each slope classes

Fig. 14. Mean sediment production in each slope class


(\*\* ) Correlation is significant at the 0.01 level (2-tailed). (\*) Correlation is significant at the 0.05 level (2-tailed).

Table 8. Pearson Correlation coefficient between different measured variables with runoff volume, sediment production and sediment concentration

Investigation of Effective Factors on Runoff Generation and

inverse relationship with sediment production of loesses.

slope. These also cause intensification of erosion and sediment production.

**19. Conclusion** 

Sediment Yield of Loess Deposits Using Rainfall Simulator 147

sediment production, Multiple regression analysis was used. For observing the phenomena of Co-linearity between variables in the extracted models, Variance Inflation Factor (VIF) was noted. In the obtained model VIF for all variables is less than the critical threshold (10) which indicates the absence of co-linearity between independent variables and that presence of all of them in the model is also meaningful. In Table 9, the summary of obtained models for produced runoff volume, sediment concentration and sediment production are shown.

Although it seems that a lot of factors are effective on runoff and sediment production of loesses, but investigation of the results show that a few number of key parameters are more important in the studied area and that other parameters have indirect effect on this matter. Slope is the most important factor in sediment and runoff production. Presence of Kaolinite decreases and the increase in the amount of silt increases sediment and runoff productions. Meyer and Harmon (1984) and Vanesland et al. (1987) also found similar results. Among chemical properties which were analyzed in this research, CEC and SAR have direct relationship and calcium cation, the amount organic matter have reverse relationships with sediment production. These are similar to a part of Hasanzade Nafuti et al. (2009), Vitharana et al. (2008) and Mahmoodabadi et al. (2009) results. Among clay mineral, smectite having week bounds between layers and being highly expandable, has direct relationship and Kaolinite being a stable clay mineral, has negative relationship with runoff and sediment production. Zhang et al. (2004) in studying loesses of China found similar relationships between erosion of loesses and the kind and amount of clay minerals. The results of regression analyses show that between the amount of runoff and sediment productions as dependent variables and independent variables with significant correlation, there is a significant relationship at %1 level. Among independent variables, only four factors: Percentage of slope, CEC, soluble Calcium cation and amount of silt, have important role in sediment production and slope and kaolinite and chlorite percentages have important role in runoff production so that these factors control %80 of sediment variations and %81 of runoff volume variations and other %20 relates to factors which are not studied in this research. Among the variables which were entered in the model, slope factor is more important so that one unit change in standard diversion of this factor produces 0.74 unit change in standard diversion of sediment production and 0.79 unit change in standard deviation of runoff volume which is because of its effect on increase of velocity, surface runoff and rain drop impact (Toy et al., 2002). Silt from the view point of size and cohesion is susceptible to erosion. CEC is effective on size and stability of soil aggregation. Calcium cation causes chemical bounds between aggregates and flocculates the grains, therefore has

In investigation of physical and chemical characteristics of loesses between surface and rill erosional features, it is found that there are not significant differences between these characteristics and the formation of special kind of erosion feature is mainly affected by slope amount and landuse type. The range lands are under overgrazing and without a plan throughout the year. This subject has caused many problems for natural reproduction of important and effected rangeland plants, the result of which being acceleration of erosion and sediment production. Construction of roads for access to cultivated lands across rangelands also has caused accelerated erosion. Also in recent years with promotion of technology, native people of the area use tractor and plough the lands in the direction of

Table 8 shows Pearson Correlation Coefficient between different measured variables with runoff volume, sediment production and sediment concentration. With statistical investigation, it was found that slope with %80 and %88 correlation coefficients, has the highest direct relationship with the amount of produced sediment and runoff, respectively, these results are similar to a part of Ribolzi et al. (2010) results. After that is the amount of silt with %53 correlation coefficient with sediment production and %54 correlation coefficient with runoff production. Materials having higher amount of silt are easily dispread and transported and are more erodible (Meyer and Harmon, 1984). The amount of clay has negative relationship with sediment and runoff production. With increase in the amount of sand, permeability is increased and lower amount of runoff is produced, in addition, despite having low adhesion and easy to be separated due to their coarser sizes, sand grains resist to transportation by runoff and produce lower amount of sediment. This result is similar to Vanesland et al. (1987) result. Vegetation cover has a negative correlation coefficient of %36 and the highest adverse relationship with sediment production and meaningful negative relationship with runoff. The reason is decrease of rainfall drop energy and velocity of surface runoff by vegetation (Yu et al. 2006). Percentage of Kaolinite has the highest negative correlation coefficient with runoff volume (%53). Other variables have weaker relationship with sediment and runoff productions. Chemical characteristics also affect runoff and sediment production so that the amount of organic matter has reverse relationship with sediment and runoff productions. In the studied area, the effect of Mg++ ion on decreasing sediment production is lower than Ca++ ion, so that it does not have meaningful correlation. Two factors of SAR and ESP, although they are not in regression models, have positive relationships with runoff and sediment production. These two are affected by other characteristics such as cations which are important in soil aggregate stability, soil infiltration and formation of surface crusts. The investigation of correlation matrix shows that none of the parameters can solely describe all observed variations in the amount of sediment.


YRun. = Total runoff volume (cm3), Sl = slope (%), Ka = kaolinite mineral (%),

Co = Chlorite mineral (%), YTurb.= turbidity(gr/lit), CEC= cation exchage capacity(c mol/kg),

Ca= Calcium cation(meq/lit), YSed.= Total sediment (gr/m2), Si= silt (%)

Table 9. Regression analyses and obtained models for runoff and sediment

For anticipating variation of sediment production based on physical and chemical properties and determination of share of each variable on explanation of the amount of sediment production, Multiple regression analysis was used. For observing the phenomena of Co-linearity between variables in the extracted models, Variance Inflation Factor (VIF) was noted. In the obtained model VIF for all variables is less than the critical threshold (10) which indicates the absence of co-linearity between independent variables and that presence of all of them in the model is also meaningful. In Table 9, the summary of obtained models for produced runoff volume, sediment concentration and sediment production are shown.
