Insert Fig. 5
Importantly, there were still variations when considering specific LUCCs. This variation was observed in different LUCCs within the same soil layer and in different layers within specific LUCC (Fig. 4, Fig. 5). On the one hand, all specific LUCCs showed significant SOC loss at different sampling depths after deforestation, except for the 0-40 cm (lnRR = 0.21, 95%CI : [-0.49, 0.14]) and 0-60 cm profiles (lnRR = -0.13, 95%CI : [-0.50, 0.46]) under the grassland transformed from the native forest (Fig. 3a). On the other hand, the response SOC storage to specific LUCC varied with the sampling depths. Generally, the lost SOC in shallow sampling (0-20 and 0-30 cm) was higher than that of deeper sampling (Fig.4 and Table 3). Overall, the lost SOC nearly 50.0% in shallow sampling (i.e., 0-20 cm, 0-30 cm) after deforestation, and less than 40.0% in deeper sampling (i.e., 0-40 cm, 0-60 cm, 0-100 cm). Similarly, there was no apparent loss of TN storage for the 0-40 cm profile after transforming from native forest to grassland. Other specific LUCCs resulted in significant lost TN storage at all sampling depths. On average, the loss efficiency of TN storage were 43.5%, 42.3%, 38.1%, 34.3% and 33.7% respectively, in the 0-20, 0-30, 0-40 and 0-60 cm profiles, showing a clear decreasing trend with depths. Therefore, these data implied that the response of SOC and TN to deforestation was pronounced in the surface layers.
In our study, we found that land restoration benefited for SOC and TN gains (Table 3, Fig. 5), but the effect magnitude showed large variation between LUCCs. Except for the slight loss after recovering from farmland to plantation, other LUCCs increased the SOC storage. Notably, the restoration from cropland to native forest significantly increased SOC for all sampling depths (Fig. 3) and the gained effect of other LUCCs decreased with sampling depths. Similarly, the increased TN storage by land restoration differed among different LUCCs. The increased of TN storage from cropland to native forest and other types (mainly indicating that grassland to native forest) was satisfactory.