Grasslands restored as part of the Conservation Reserve Program (CRP) span spatial and temporal scales and therefore represent model systems useful in improving our understanding of mechanisms underlying ecosystem recovery from disturbance. Ecosystem recovery during restoration has been demonstrated in multiple chronosequence studies in C₄ grass-dominated grasslands. However, the potential role of abiotic factors in influencing recovery trajectories has received considerably less attention. To investigate the effect of soil texture on ecosystem recovery of restored grasslands, we compared soil carbon (C) and nitrogen (N) pools and fluxes in two chronosequences on contrasting soil types in eastern Nebraska: silty clay loam (SCL) (3 agricultural fields, 3 native prairies, and 19 restored grasslands ranging from 2-18 years following restoration) and loamy fine sand (LFS) (3 agricultural fields, 3 native prairies, and 21 restored grasslands ranging from 2-18 years following restoration).

Results/Conclusions

Total soil C in SCL native prairies (3724 ± 193 g/m²) exceeded that of LFS sites (2240 ± 513 g/m²), and recovery of most belowground properties proceeded towards native prairie in SCL soils. Within the SCL chronosequence, total soil C (24.67 g/ m²/yr, r²=0.34, P<0.001) and C/N ratios (0.09/yr, r²=0.55, P<0.0001) increased linearly with years restored. Also within the SCL chronosequence, both microbial biomass C (MBC: 3.05 g/m²/yr, r²=0.39, P<0.001) and MBN (fall only; 0.26 g/ m²/yr, r²=0.28, P=0.002) increased over time. Potential net N mineralization (PNM) rates decreased over the SCL soil chronosequence as well (-10.28 mg/m²/yr, r²=0.45, P<0.0001), reflecting N-limitation typically found in native prairie systems. In contrast, recovery across the LFS soil chronosequence was limited. Although C/N ratios increased with years restored (0.07/yr, r²=0.17, P=0.083), neither total soil C (r²=0.04, P=0.406) nor total soil N (r²<0.01, P=0.799) showed evidence of recovery along the LFS chronosequence. Microbial C also showed no recovery over time (MBC: r²=0.12, P=0.131), but MBN did increase (0.16 g/m²/yr, r²=0.56, P<0.0001) across the LFS chronosequence. Rates of PNM decreased with years restored in LFS soil (-2.74 mg/m²/yr, r²=0.36, P=0.004), but at a slower rate than SCL soils. Our results demonstrate that abiotic site characteristics can substantially impact recovery rates in restored grasslands. In this case, recovery of pools and fluxes of soil C and N were greatly constrained in sites with sandy soils. This suggests a much longer recovery time, which underscores the need for conservation of remaining in-tact grasslands that developed on predominantly sandy soils.