Uses of Ecosystem Modeling for Range Management

Xuejun Dong, Paul Nyren, Bob Patton, Brian Kreft, and Anne Nyren, NDSU Central Grasslands Research Extension Center, Streeter, ND


Table of Contents

List of Tables and Figures



In the Northern Great Plains, the proper management of native grasslands is very important for both the present-day economy and the continued use of rangeland by future generations. The rangelands of this area have a unique origin and ecological regime and a science-based understanding plays an important role in its better management. At the Central Grasslands Research Extension Center (CGREC), several scientific projects are being conducted or will soon be initiated for a better understanding of the range production processes. Ecosystem modeling is one of them. Here the term “ecosystem” refers to the workings of the plants (grasses, forbs, shrubs) as influenced by various other biotic (animals, microorganisms, etc.) and abiotic (solar radiation, soil resources, air, etc.) factors or agents, and the term “modeling” refers to the use of the computer and mathematics to quantitatively describe the various processes within the grassland ecosystem, and to predict their future changes in relation to grassland productivity. A “model” of a real ecosystem usually captures its essential features but ignores others that are assumed to be trivial in terms of particular objectives of the research. In this short introduction, we will highlight some interesting properties of ecosystems, the methods we are using for modeling at the Central Grasslands Research Extension Center and their implications to range management.

Ecosystem Highlights

Flow networks: Figure 1 is a simplified picture of a grassland ecosystem showing the flow networks within it. The flows can be materials, energy or information that are directional and can usually be quantified by experimentation. The following points deserve special attention: (a) The ultimate push of the flows comes from the solar energy. (b) There is always an efficiency associated with material turnover in the ecosystem. About 50% of plant dry mass is carbon and about 30% of the carbon fixed in gross photosynthesis is respired by the plant. In animals, about 65% of the carbon intake is respired (only 0.2% is retained in the animal body). The ultimate purpose of many management measures is to increase the carbon retention in plant and animal bodies (e.g. increasing water use efficiency in plants and using low-stress methods in livestock handling). (c) Fertilization contributes directly to the soil solutes pool; while haying and grazing remove plant material. Animal urine and feces are important inputs to soil mineral nitrogen (about 99% of the nitrogen consumed by animal is returned to the soil through excreta and only 1% retained in the animal body).

Dynamics: Another interesting view of an ecosystem is its dynamic nature. The term “dynamic” is used here to describe something that changes over time (opposite to static). It’s quite straightforward to attribute ecosystem dynamics to the daily and seasonal variations in climate and weather. However, we will show here that there is another source of dynamics that originates from the inside of an ecosystem itself! That is the drastically varied response times for many entities within an ecosystem (Table 1). The fastest entity in Table 1 changes in a tenth of a second, while the slowest changes in hundreds of years. The following points deserve noting: (a) Given optimal conditions, many properties of an ecosystem are mostly determined by the slow pools. However, when the environments are away from optimum condition, many fast pools may have a big influence to the whole system behavior (for example, a study of plant water use mechanisms becomes more important in drought years). (b) When many entities with contrasted response times are present each with significant influence to the ecosystem behavior (especially under actual climate situations), it’s harder for the ecosystem to obtain an equilibrium condition. (c) The long response time of the largest soil organic matter (SOM) pool implies that if the SOM is disturbed (e.g. by frequent plowing), it is very hard to rebuild the SOM pool to its original level. This is one of the reasons why in areas like the Missouri Coteau grazing is a recommended grassland use.

Hierarchical structures: An ecosystem is composed of different levels of organizations that contribute to the diversity of the biological world. In a plant ecosystem, we can list many levels from small to large: cells, organs (root, leaves), plant, ecosystem, etc. Each level has some kind of independence. For example, a piece of plant leaf (of one inch long, one inch wide and one-hundredth inch thick), though very light, contains about 1 million cells, each functioning more or less independently, but communicating with neighboring cells through various ways. Processes at a lower level are usually considered to be mechanisms for a higher level, while processes at a higher level usually set constraints to those at lower levels.

Research Method

Since a model of an ecosystem is a simplified representation of it, the key element of our research method is linking an ecosystem model to field measurements. At the CGREC, several field experiments are in progress, including a grazing intensity study (Patton, et al. 1998), a drought study with automatic rain-out-shelter (Kirby, et al. 1999), a fertilizing study (Volk, et al. 2000), a photosynthetic plant water-use study (Dong, et al. 2001), etc. In the context of ecosystem hierarchical structure, these studies either deal with the physiological (organ, plant) level or the ecosystem level. Data obtained from physiological-level studies are mostly some rate variables and can be used for the estimation of some model parameters governing various fluxes in the ecosystem; data from ecosystem-level studies are mostly some status variables and can be used for model testing, or as reference properties of the grassland ecosystem.

Ecosystem Modeling and Range Management

Different soil and habitat conditions: Field rangeland experiments are conducted at a particular site. And due to the long-term nature of many ecological experiments, it is hard to test as many possibilities as required by producers. Producers may be concerned that the research conducted at research centers is of limited usefulness to them because the site conditions (soil physical, chemical properties etc.) of the research centers may be different from their rangeland. It is possible to overcome this problem with the aid of ecosystem modeling. In an ecosystem model, various factors and processes are described quantitatively and the manipulation of them is relatively easy. When the model is confronted more and more by experimental data, its prediction ability increases. Then, changing to a new location may just involve some minor adjustments to a few model parameters and/or some initial conditions. Also, it is possible to try some other “what-if” possibilities with the aid of computer experimentation. These possibilities may include ones that were never tested, or are impossible to test by the field experiments but are relevant to management.

Fertilizers: Fertilizers, if used appropriately, can significantly increase rangeland primary production (Lorenz, 1970; Nyren, 1979; Volk, et al. 2000). One important issue is how to effectively use fertilizers to achieve a production goal. Field experiments have shown that fertilizers no longer boost production beyond some appropriate levels (frequency and/or intensity) of use. In a particular range situation, a solution of higher precision could be obtained with the aid of ecosystem modeling. Another problem associated with the use of fertilizers for range improvement is the appropriate interpretation of the variability among the results of different experiments. Variability happens especially when environments are sub-optimal, or when more factors are included in the experiment. It may be attributed to fertilizer types (See Volk, et al. 2000 for discussions on the ecological implications of using fast release vs. slow release fertilizers), frequency and amount of rainfall, soil physical properties and soil water flows, plant nutrient uptake rate, grazing pressure, and so on. When many factors and processes are involved, the framework of ecosystem concepts (its flow networks, dynamics and hierarchical structures) and the method of ecosystem modeling become important in the understanding of the related causal relationships, and in the interpretation of the experimental variability and its management implications.

Periodic drought: Climatic drought has a fundamental control over the productivity of grasslands in the Northern Great Plains (Biondini et al. 1998; Kirby et al. 1999). In the past one hundred years, climatic records show that drought occurred in this area in a more or less cyclic way. And thus we have the so-called wet and dry periods. Using a physiologically-based ecosystem model which is capable of predicting for hundreds of years, it is possible to provide some strategic guidelines to range management for dealing with different possibilities of the future climate, given the particular climatic trajectories the grasslands ecosystem has experienced in the past. The drought-caused reduction in plant growth potential and post-drought recovery can be complex (drought has a fundamental influence on various processes in the soil-plant system). We will use the rain-out-shelter facility to quantify major physiological parameters of plants prior, during and after drought stress. This data will be helpful for the modeling effort of grassland drought responses.


Plant species consideration: To emphasize the ecosystem complexity we deliberately detailed the soil processes and simplified plant processes in Figure1. However, the ecosystem model we are using has a strong emphasis on plant physiology. Specifically, the warm-season and cool-season plants will be treated separately. These two types of plants, with their contrasted physiology and environmental responses, are very important in both grasslands and croplands. In the northern part of the Great Plains, there are some observations that the grasslands tend to become cool-season dominated. Also, numerous greenhouse experiments have shown that many cool-season grasses will be favored over warm-seasons in the predicted increase of atmospheric CO2. However, the field success of these plants depends on many factors (just consider the complexity and dynamics of ecosystem). If, for example, decades later the grasslands of this area became extremely dominated by cool-season grasses. Then, the climate scenario of cold winter plus hot summer would bring about a shortage of herbage production. This would have a drastic affect on the beef industry. What is the long-term ecosystem consequence of similar hypothetical ecological changes and what is a better management strategy in similar situations? As discussed by Volk et al. (2000), the long-term balance between the cool-season and warm-season grasses could be affected by the different using of fertilizers (slow-release vs. fast-release). The method of ecosystem modeling is useful for the further analysis of the similar problems. Last, when genetically modified plant species are used at a considerable scale, ecosystem modeling could be a useful tool to assess the possible consequences.

Our use of ecosystem modeling is mostly to predict short-term and long-term rangeland dynamics and its management implications, and to reveal cost-effective ways to increase grassland production potential. Processes in the soil-plant system are described in detail, while animal metabolism is very simplified and animal reproduction and range product marketing are not included at all. To understand the whole picture of the grassland system (soil-plant relations, beef production, economics), cooperation between different disciplines is necessary.
Table of contents

NDSU Central Grasslands Research Extension Center
Home | Livestock Research | Economics Research | Range Research | 2001 Annual Report