Differential response of mineral-associated organic matter in tropical soils formed in volcanic ashes and marine Tertiary sediment to treatment with HCl, NaOCl, and Na4P2O7

Quantitative knowledge of the amount and stability of soil organic matter (SOM) is necessary to understand and predict the role of soils in the global carbon cycle. At present little is known about the influence of soil type on the storage and stability of SOM, especially in the tropics. We compared the amount of mineral-associated SOM resistant to different chemical treatments in soils of different parent material and mineralogical composition (volcanic ashes – dominated by short-range-order aluminosilicates and marine Tertiary sediments – dominated by smectite) in the humid tropics of Northwest Ecuador. Using 13C isotope analyses we traced the origin of soil organic carbon (SOC) in mineral-associated soil fractions resistant to treatment with HCl, NaOCl, and Na4P2O7 under pasture (C4) and secondary forest (C3). Prior to chemical treatments, particulate organic matter was removed by density fractionation (cut-off: 1.6 g cm−3). Our results show that: (1) independent of soil mineralogical composition, about 45% of mineral-associated SOC was resistant to acid hydrolysis, suggesting a comparable SOM composition for the investigated soils; (2) oxidation by NaOCl isolated a SOM fraction with enhanced stability of mineral-bound SOM in soils developed from volcanic ashes; while Na4P2O7 extracted more SOC, indicating the importance of Al-humus complexes in these soils; and (3) recently incorporated SOM was not stabilized after land use change in soils developed from volcanic ashes but was partly stabilized in soils rich in smectites. Together these results show that the employed methods were not able to isolate a SOM fraction which is protected against microbial decay under field conditions and that the outcome of these methods is sensitive to soil type which makes interpretation challenging and generalisations to other soils types or climates impossible.