Formation of mineral-organic associations is a key process in the global carbon cycle. Recent concepts propose litter quality-controlled microbial assimilation and direct sorption processes as main factors in transferring carbon from plant litter into mineral-organic associations. We explored the pathways of the formation of mineral-associated organic matter (MOM) in soil profiles along a 120-ky ecosystem gradient that developed under humid climate from the retreating Franz Josef Glacier in New Zealand. We determined the stocks of particulate and mineral-associated carbon, the isotope signature and microbial decomposability of organic matter, and plant and microbial biomarkers (lignin phenols, amino sugars and acids) in MOM. Results revealed that litter quality had little effect on the accumulation of mineral-associated carbon and that plant-derived carbon bypassed microbial assimilation at all soil depths. Seemingly, MOM forms by sorption of microbial as well as plant-derived compounds to minerals. The MOM in carbon-saturated topsoil was characterized by the steady exchange of older for recent carbon, while subsoil MOM arises from retention of organic matter transported with percolating water. Overall, MOM formation is not monocausal but involves various mechanisms and processes, with reactive minerals being effective filters capable of erasing chemical differences in organic matter inputs.

In soils, a large portion of organic nitrogen (ON) is associated with minerals and thus, possibly stabilized against biological decay. We therefore tested if mineral-associated N is an important N source for soil microorganisms, and which soil parameters control its bioavailability. Microcosm experiments with mineral-associated organic matter, obtained as heavy fraction (HF) via density fractionation, and bulk soil from mineral topsoil of the Franz Josef chronosequence were conducted for 125 days. We examined the effects of O2 status, soil age (differences in mineralogical properties), as well as cellulose and phosphate additions on the turnover of mineral-associated N. Using a combination of activity measurements and quantitative PCR, microbial N transformation rates and abundances of N-related functional genes (amoA, narG, chiA) were determined. Similar or higher values for microbial N cycling rates and N-related functional abundances in the HF compared to bulk soil indicated that mineral-associated N provides an important bioavailable N source for soil microorganism. The turnover of mineral-associated N was mainly controlled by the O2 status. Besides, soil mineralogical properties significantly affected microbial N cycling and related gene abundances with the effect depending on the N substrate type (ON, NH4+ or NO3−). In contrast, cellulose or phosphate addition hardly enhanced microbial utilization of mineral-associated N. The results of our microcosm study indicate that mineral-associated N is highly bioavailable in mineral topsoils, but effects of the mineral phase differ between N cycling processes.

Along a long-term ecosystem development gradient, soil nutrient contents and mineralogical properties change, therefore probably altering soil microbial communities. However, knowledge about the dynamics of soil microbial communities during long-term ecosystem development including progressive and retrogressive stages is limited, especially in mineral soils. Therefore, microbial abundances (quantitative PCR) and community composition (pyrosequencing) as well as their controlling soil properties were investigated in soil depth profiles along the 120,000 years old Franz Josef chronosequence (New Zealand). Additionally, in a microcosm incubation experiment the effects of particular soil properties, i.e. soil age, soil organic matter fraction (mineral-associated vs. particulate), O2 status, and carbon and phosphorus additions, on microbial abundances (quantitative PCR) and community patterns (T-RFLP) were analyzed. The archaeal to bacterial abundance ratio not only increased with soil depth but also with soil age along the chronosequence, coinciding with mineralogical changes and increasing phosphorus limitation. Results of the incubation experiment indicated that archaeal abundances were less impacted by the tested soil parameters compared to Bacteria suggesting that Archaea may better cope with mineral-induced substrate restrictions in subsoils and older soils. Instead, archaeal communities showed a soil age-related compositional shift with the Bathyarchaeota, that were frequently detected in nutrient-poor, low-energy environments, being dominant at the oldest site. However, bacterial communities remained stable with ongoing soil development. In contrast to the abundances, the archaeal compositional shift was associated with the mineralogical gradient. Our study revealed, that archaeal and bacterial communities in whole soil profiles are differently affected by long-term soil development with archaeal communities probably being better adapted to subsoil conditions, especially in nutrient-depleted old soils.

During long-term ecosystem development, both soil mineralogical composition and nutrient contents change, thus possibly altering microbial nutrient cycling by constraining substrate accessibility. In addressing the mineral impact on nitrogen (N) and phosphorus (P) cycling, we determined microbial abundances, activities of N-hydrolyzing (aminopeptidases, protease, urease) and P-hydrolyzing (phosphatase) enzymes and the potential substrate availability as well as their physicochemical and mineralogical controls in whole soil profiles along the 120 kyr-old Franz Josef chronosequence (New Zealand). Pedogenic soil iron (Fe) and aluminum (Al) resided initially (<1 kyrs) in metal-humus complexes, changed to poorly crystalline Fe and Al at intermediate-aged sites (1-12 kyrs) and into dominance of clay and crystalline Fe oxides at the oldest site. Despite this, organic C (OC) and organic N (ON) stocks increased only slightly with soil age, whereas organic P (OP) stocks decreased continuously. In organic layers, enzyme activities were mainly regulated by ON and OP concentrations, whereas in mineral soils, mineral enzyme relations were more complex and included both, direct and indirect effects. Protease, urease, and phosphatase activities were inhibited by mineral interactions, especially with poorly crystalline Fe and Al oxides, whereas aminopeptidases were less affected by mineralogical properties. On a pedon basis, most N-hydrolyzing enzyme activities per ON stocks responded negatively to increasing stocks of poorly crystalline Fe and Al minerals, but were also affected by the C:N ratio of labile organic substrates. Profile-based phosphatase activities per OP stock were highest at the oldest sites having the largest stocks of clay and crystalline Fe oxides. Overall, our study indicates that long-term mineral changes create distinct patterns of nutrient accumulation and N- and P-enzyme activities at both horizon and pedon scale, with a variable extent of the mineralogical effect for the different N-hydrolyzing enzymes. (C) 2013 Elsevier Ltd. All rights reserved.