591 CONCEPTS & SYNTHESIS EMPHASIZING NEW IDEAS TO STIMULATE RESEARCH IN ECOLOGY Ecology, 85(3), 2004, pp. 591���602 q 2004 by the Ecological Society of America NITROGEN MINERALIZATION: CHALLENGES OF A CHANGING PARADIGM JOSHUA P. SCHIMEL1,3 AND JENNIFER BENNETT2 1Department of Ecology, Evolution, and Marine Biology, University of California, Santa Barbara, California 93106 USA 2Forest Nutrition Coop, North Carolina State University, Raleigh, North Carolina 27695-8008 USA Abstract. Until recently, the common view of the terrestrial nitrogen cycle had been driven by two core assumptions���plants use only inorganic N and they compete poorly against soil microbes for N. Thus, plants were thought to use N that microbes ������left over,������ allowing the N cycle to be divided cleanly into two pieces���the microbial decomposition side and the plant uptake and use side. These were linked by the process of net mineral- ization. Over the last decade, research has changed these views. N cycling is now seen as being driven by the depolymerization of N-containing polymers by microbial (including mycorrhizal) extracellular enzymes. This releases organic N-containing monomers that may be used by either plants or microbes. However, a complete new conceptual model of the soil N cycle needs to incorporate recent research on plant���microbe competition and mi- crosite processes to explain the dynamics of N across the wide range of N availability found in terrestrial ecosystems. We discuss the evolution of thinking about the soil N cycle, propose a new integrated conceptual model that explains how N cycling changes as eco- system N availability changes, and discuss methodological issues raised by the changing paradigm of terrestrial N cycling. Key words: microsites nitrogen availability nitrogen mineralization N-mineralization para- digm, evolving plant uptake soil N cycle. DEVELOPMENT OF THE ������CLASSICAL������ PARADIGM OF N CYCLING Since the late 1800s, N mineralization has been the perceived center point of the soil N cycle and the pro- cess that controls N availability to plants (Russell 1912, Waksman 1932, Harmsen and Van Schreven 1955, Aber and Melillo 2001). This view grew from two par- allel and complimentary threads in the development of our understanding of plant���soil interactions. The first thread was the adoption of the mineral-nutrition theory of plant nutrition: with the writings of Liebig (1842), it became the widely accepted view that plants use only inorganic materials for their nutrition. Despite reports as early as the late 1800s that some plants actually can use organic N (Waksman 1932, Harmsen and Van Schreven 1955), the mineral nutrition theory remained effectively unchallenged for 150 years, as suggested by Black (1993:383): Manuscript received 17 January 2003 revised 29 July 2003 accepted 3 August 2003. Corresponding Editor: P. M. Groffman. 3 E-mail: Schimel@lifesci.ucsb.edu A small amount of organic nitrogen is found in the soil solution, but plants are not known to take up any significant part of it, and must depend upon the inorganic forms that are released when microor- ganisms decompose the compounds containing the organic forms. The second key thread in developing the current view of the nitrogen cycle was the recognition that decom- position is a microbial process with NH41 as a waste product, as described by Waksman (1932:444): As long as there is free available energy, in excess of the available nutrients, there will be only a min- imum accumulation of available plant food. When the energy approaches exhaustion ammonia (or ni- trate) begins to accumulate. . . Since the microor- ganisms are unable to assimilate it, due to the ab- sence of sufficient available energy material, it is left in the soil for the use of higher plants. These ideas framed the two core assumptions of N cycling studies that established mineralization as the perceived center point of the N cycle, and that most researchers have used for the last century when ex-

Concepts & Synthesis 592 JOSHUA P. SCHIMEL AND JENNIFER BENNETT Ecology, Vol. 85, No. 3 amining soil N processes: (1) Plants only use inorganic N. (2) Plants are poor competitors for available soil N relative to microbes. They ������lose������ almost all compe- tition events, and therefore effectively only have access to N that is left over after microbial N demands are met. More specifically and importantly, these assump- tions established net mineralization as the key step in soil N cycling and as defining the amount of N available either for plant uptake or loss from the ecosystem (Vi- tousek et al. 1979). This tenet is repeatedly indicated by standard references of the 1970s and 1980s: Part of the mineral nitrogen produced (Nmin), pref- erably ammonium, is consumed by the soil micro- organisms (Jansson, 1958). As they are more suc- cessful than the higher plants in competing for Nmin (Bartholomew and Clark, 1950 Jansson, 1958 Zo ��ttl, 1960a), of the total Nmin production (gross mineralization), the higher plants can use only that part (net mineralization) which exceeds the micro- bial demand. ���Runge (1971:191) (paraphrased in Melillo [1981:437])4 In unfertilized grasslands, uptake capabilities of mi- croorganisms and plants together generally exceed the mineralization potentials of systems. . . . In such an environment, plants and microorganisms are in intense competition, often one or both classes of or- ganisms will not meet their ������demands,������ and growth will be limited. During these periods of competition, microorganisms probably have the competitive ad- vantage because of their intimate association with the substrate. ���Woodmansee et al. (1981:4491) This thinking about the role of net mineralization naturally led to the development of net mineralization assays as the standard tool for measuring plant-avail- able N. This was an important development in eco- system science because developing theory saw nutrient availability and uptake as critical factors governing ecosystem development (Odum 1969, Vitousek and Re- iners 1975), function (Chapin 1980, Gutschick 1981, Vitousek 1982), and response to disturbance (Vitousek et al. 1979). Further development was considered to be limited by the lack of an independent measurement of soil N availability: ������Ideally, nitrogen availability should be measured independently of the amount of nitrogen in litterfall, but direct measures of nitrogen availability in forest soils are not now possible������ (Vi- tousek 1982:564). Thus, the application of techniques such as the ������buried bag������ (Eno 1960) and the analogous but more sophisticated ������resin core������ (DiStefano and Gholz 1986) incubations became widely used, provid- 4 The citations within this quotation were not consulted by the authors in the preparation of this paper, but are included in the Literature Cited for the reader���s convenience. ing ecosystem ecologists with what was felt to be the best, broadly applicable, in situ assay of N available for plant uptake. For a number of years, such in situ net mineralization assays were considered an adequate measure of plant-available N: Annual N uptake by vegetation at each site was as- sumed to be annual net mineralization in the 0���10 cm soil zone plus mineral N in precipitation minus mineral N leached from the rooting zone. ���Nadelhoffer et al. (1983:15) An important feature of recent in situ methods [of measuring net mineralization] is that, with appro- priate assumptions, uptake of inorganic-N by forests can be calculated. ���Adams et al. (1989:423) Consequently, the main concerns with net mineral- ization assays were not focused on the underlying the- ory, but with artifacts associated with physically dis- rupting the soil, cutting roots and thereby changing C availability (Adams et al. 1989), and altering soil mois- ture (Tietema et al. 1992). Over the course of the 1990s, however, thinking about net mineralization became more refined with the developing awareness of the technical limitations of the assays and of the complex dynamics of gross min- eralization/immobilization reactions (Schimel et al. 1989, Tietema and Wessel 1992, Hart et al. 1994a). Studies showing substantial rates of gross minerali- zation and nitrification in systems where little NH41 and NO32 accumulates during net mineralization assays particularly highlight the limitations of net rates as measures of N cycling dynamics (Davidson et al. 1992, Hart et al. 1994b, Neill et al. 1999). As a result, few papers now use net mineralization assays as a direct measure of plant-available N for whole-ecosystem bud- geting purposes. Instead, net N mineralization is dis- cussed as an ������index������ rather than as a ������measure������ of plant-available N: Net N mineralization provides an index of plant available N in many systems (Nadelhoffer et al. 1983), but does not reflect the total amount of N cycling between organic matter and soil inorganic N. ���Neill et al. (1999:567) The net accumulation of inorganic N in the absence of plant roots is thought to provide a good index of N availability to plants. ���Hart et al. (1994a:999) The growing concern with net-mineralization assays that caused the shift in thought and language has not, however, eliminated its use as a measurement tool. Rather, net mineralization remains useful for some ap- plications as a tool for assessing available N (e.g.,