Keywords: Global biogeochemical nitrogen cycle, Ecosystem model, Soil Nitrogen, Storage, Fluxes, Geographical information system

Abstract
To quantitatively evaluate the cause-and-effect relationships of human land-use changes to terrestrial ecosystems requires preliminary analyses of potential nitrogen (N) fluxes and its storage in natural soil. This paper uses a mechanistically based ecosystem simulation model of global N cycle highly integrated with carbon (C) to aim at satisfying the demand. This model treats the global environment in a total of 60156 grid cells with 0.5° in its latitude and longitude, and uses information on soils, vegetation, and climate variables such as temperature, precipitation, and solar radiation as independent driving forces to run the model. Our model-based estimate of potential N storage in global natural soil is 350Pg, which shows good agreement with that of extant reported as ranging from 70Pg to 820Pg, with a few intermediate estimates of 170Pg, 760Pg, 175Pg, and 300Pg. Of which 350Pg in soil, 24Pg is in inorganic forms as ammonium and nitrate and 326Pg in organic forms as detritus and humus. More than 90% of the soil N represents as in organic forms, which agrees very well on various field researches. Estimates of mean N storage in each type of ecosystem range from 50-860g/m² in inorganic forms and 1480-7550g/m ² in organic forms, in which the tropical forest soil shows the least storage due to its higher decomposition rates. Its global distribution based on model’s estimates is very similar to that made by Wlilfred (1985). Nevertheless additional work on model validations are still needed, this model has provided us a better view of the global N storage.

Introduction
Nitrogen (N) is both an essential nutrient and a major pollutant in the terrestrial ecosystems. Human land-use changes such as fertilization, cultivation, and deforestation have significantly disturbed the N balance in natural ecosystem, that result in various environmental problems at regional to global scale (Daniel, 1992, Nyle, 1998). Ecosystem models that can integrate the information derived from site-specific researches or empirical studies or remote-sensing imagery, and extrapolate this information to other locations and through time, have been widely recognized as useful research tool for evaluating the cause-and-effect relationships of human land-use changes to terrestrial ecosystems. Many attempts toward this end have been seen in recent decades. However, most of them are from the standpoint of carbon (C) cycle model, and still very few from the standpoint of nitrogen (N) cycle are found (Christopher et al., 1993, 1997; Robert et al., 1994). Terrestrial Ecosystem Model developed by Melillo J.M. group (McGuire et al., 1992; Melillo et al., 1993; Raich et al., 1991) is the most valuable to be mentioned among several other N cycles (Michel et al., 1997; Rastetter et al., 1991). But this model mainly focused on C cycles and simplified the input and output fluxes of N by incorporating them into the model as constants. In our previous paper (Binle et al., 1999), we have fully described a global biogeochemical N cycle model, which has been developed as a highly aggregated, process-based simulation model that includes almost all of the important processes governing N cycles in natural ecosystem. In this paper the model has been used as a research tool to get insight the potential N fluxes and storage in natural soil, which is very necessary for the attempt of quantitatively evaluatingthe potential impacts of land-use and climate changes.

Methods and Data

The Model
The amount of N stored in soil is related to climate through biotic process associated with productivityof vegetation and decomposition of organic matter. Other factors, particularlyrain-fall input, dry deposition input, nitrogen fixation and losses of inorganic nitrogen due to leaching contribute to the variability of N storage. The model used here incorporated all of the remarked factors, its model structure is depicted in Fig. 1, while the components and processes involved in each compartment are summarized in Table 1. The model treats the global atmosphere as a one-well-mixed reservoir, while treats the terrestrial biosphere reservoir in a total of 60156 grid cells with a spatial resolution of 0.5° in its latitude and longitude. Each grid cell is fixed to be one of the five types of ecosystems based on the model output of our previous global C model study (Motoyuki et al., 1993; Naohiro et al., 1994). Model validations and sensitivity analyses carried in our previous paper (Binle et al., 1999) confirm that our mechanistically process-based model can simulate verywell the N cycles in natural ecosystems.

Fig. 1 Model structure of global biogeochemical nitrogen cycle coupled with carbon cycles.


Table 1 Definitions of reservoirs, compartments, components, and processes involved in model.

Reservoirs	Compartments	Components	Processes
Atmosphere	One-well mixed pool	Nitrous oxide	Biomass burning Lightning
Terrestrial	0.5 ° grid-based	Eight state variables:	Thirteen fluxes:
biosphere	Three compartments in vertical	Nitrogen in vegetation	Photosynthesis
	Vegetation: Leaf, Trunk and Root	Carbon in vegetation	Plant respiration
	Soil organic: Detritus and Humus	Nitrogen in detritus	Litter-fall
	Soil inorganic: Ammonium and Nitrate	Carbon in detritus	Nitrogen fixation
	Five ecosystems in horizontal	Nitrogen in humus	Detritus decomposition
	Tropical forest	Carbon in humus	Huminification
	Temperate forest	Ammonium in soil	Humus carbonization
	Boreal forest	Nitrate in soil	Plant uptake
	Grass/Agricultural		Nitrification
	No vegetation (desert and ice)		Humus decomposition
			Ammonia volatilization
			Denitrification
			Nitrate leaching