Simulating Grazing Effects on Soil Organic Carbon Dynamics in Semi-arid Rangelands (Southern Iran)

Grazing is one of the main causes of rangeland degradation worldwide, due to the 23 effects of overgrazing on vegetation cover and biodiversity. But few data are available on 24 the effect of grazing intensity on the dynamics of soil organic carbon (SOC) and soil labile 25 organic carbon (SLOC). So far, very few studies have addressed the modeling of SOC 26 dynamics under different grazing intensities, and SLOC dynamics has not been modeled 27 yet. In this study, we used the CENTURY model to select the most effective grazing 28 management in terms of carbon sequestration (SOC and SLOC stocks) in semi-arid 29 rangelands of Southern Iran. The effect of four different scenarios of grazing intensity was 30 simulated: no grazing, light grazing (LG), moderate grazing (MG), and heavy grazing 31 (HG). The results of long-term model simulations (2015-2100), indicated that SOC stocks 32 will change by 2.7, 1.7, -23.4, and -24.6% in the scenarios of exclusion, LG, MG, and HG 33 respectively compared to 2014. With increasing grazing intensities, SLOC stocks in LG, 34 MG, and HG scenarios significantly decreased compared to the no grazing scenario by 35 26.1, 59.6, and 70%, respectively. Thus, this study suggests recommending light grazing 36 management for semi-arid rangelands of Iran and also SLOC as a suitable index for studying the effect of grazing on soil carbon.


41
Soils have a prominent role in maintaining the balance of the global carbon cycle (La,42 2008), and any small change in soil organic carbon (SOC) has high impacts on the 43 concentration of CO2 in the atmosphere (Smith et al., 2008;Muñoz-Rojas et al., 2015). 44 Rangelands have a high potential to sequester the atmospheric CO2 in the soil due to their 45 prevalence in about 50% of land area worldwide, and globally can store up to 30% of SOC 46 (Derner and Schuman, 2007). In particular, one of the most effective factors in SOC storage 47 is grazing management (Mcsherry and Ritchie, 2013;Waters et al., 2016). Grazing is one 48 of the main causes of rangeland degradation particularly in arid and semi-arid 49 environments, due to the effects of overgrazing on vegetation cover, biodiversity of plant 50 species, unpalatable species, and livestock trampling that enhance soil loss by erosion, 51 reduce the production potential of rangelands, and negatively affect SOC, SOC pools and  (Geng et al., 2009). SLOC has a relatively short turnover time and has 69 shown higher sensitivity to management practices when compared to the total SOC stock. 70 Thus, this pool has been suggested as suitable and sensitive indicator to study the effect of 71 grazing on SOC (Soon et al., 2007;Cao et al., 2013). 72 In recent years, simulation models have been recognized as effective tools for decision-73 making and ecosystems management in relation to soil carbon sequestration. The 74 CENTURY model is a process-based ecological model to simulate SOC dynamics, 75 integrating the effects of climate, soil driving variables and management in different 76 ecosystems (croplands, grasslands, forests and savannas) on soil fertility parameters and 77 water dynamics (Parton et al., 1987). The model includes specific options to simulate the 78 effect of grazing on plant production and soil carbon. Shiraz University, these pastures were not grazed after the revolution. Our hypothesis was 86 that no grazing and light grazing would not have a negative effect on soil carbon stock, and 87 with increasing grazing intensity in the long term, soil organic carbon stock and the soil 88 carbon sequestration rate would decrease in these rangelands. If the results of the studies 89 showed us that a type of grazing intensity management cannot have a negative effect on 90 soil carbon, we would be able to propose this type of grazing management to the school 91 authorities, which will not reduce the soil organic carbon stock (soil health), and it can also 92 provide animal husbandry and livestock products. Therefore, the aims of this study are: i)   The rangelands of Bajgah are classified as semi-arid, and grasses are mainly C3 type.   Particle-size distribution, soil reaction, soil organic carbon, and total soil nitrogen were 162 determined by the Hydrometry method (Bouyoucos, 1962), pH meter, Walkley and Black 163 method (Walkley and Black, 1934), and Kjeldahl method (Bremner et al., 1982) 164 respectively. To determine the soil bulk density, two soil core samples (in addition to 165 routine soil samples) were collected in each sampling point using the Core method (Blake 166 and Hartge, 1986). Finally, SOC stock was calculated in the 0-20 cm layer using Eq. (1):  period. At the equilibrium state, the fractions of slow, active, and passive organic carbon 241 pools were 50%, 3.8%, and 46.2% of the total SOC respectively.

242
There was a significant linear relationship (R 2 =0.86) between measured and simulated 243 SOC stocks (Figure 4), and the correlation coefficient (r) was 0.93 ( in the Bajgah rangelands ( Figure 4).

Changes of SOC stocks under grazing 250
Long-term changes of SOC stocks in response to grazing until 2100 showed significant 251 differences (p<0.01), and the minimum (2883.4 g m -2 ) and maximum (3436.2 g m -2 ) SOC 252 stocks were observed in the heavy and no grazing scenarios (Table 3). Figure 5 shows the  (Table 3).

261
At higher grazing intensities, SOC stocks decreased significantly in the two scenarios 262 of moderate and heavy grazing in comparison with the no grazing and light grazing 263 scenarios (Table 3 and Figure 5) (Table 4).

276
The results of validation indicated that the CENTURY model was able to simulate the 277 changes of SOC (Table 2), and can be applied for the long-term simulation of the SOC

301
In the Bajgah rangelands, long-term simulations indicated that the light grazing 302 intensity presents increased SOC stocks compared to moderate and heavy grazing ( Figure   303 5), with not significant difference compared with the no grazing scenario (Table 3). Wang  The results of the simulation indicated decreased levels of soil labile organic carbon 351 (SLOC) with increased grazing intensity in the Bajgah rangelands ( Figure 6). Cao et al.

352
(2013) also stated that with increased grazing intensity, the SLOC stocks would reduce.

353
The SLOC stocks decreased by 59.6 and 70% under moderate and heavy grazing scenarios 354 respectively, in comparison with the no grazing scenario (Table 4)

358
Although the amount of SOC stocks had no significant difference between the light 359 grazing and no grazing scenarios, as well as between moderate and heavy grazing scenarios 360 (Table 3), there was a significant difference (p<0.01) among the levels of SLOC stocks in 361 all grazing scenarios (Table 4). In addition, the annual changes of the SLOC pools in 362 comparison with the annual variations of SOC stocks showed that SLOC is more sensitive 363 to grazing and it responds much faster to grazing than SOC ( Figure 5 and 6)     The Century model owchart.

Figure 3
Change of carbon pools during the equilibrium state process.

Figure 4
Measured and simulated SOC Stock in compared with 1:1 line. Vertical bars indicate the difference between measured and simulated values. EF compares simulations or predictions and observations on an average level, and ranges from − 1 to 1, with best performance at EF = 1.