![]()
Analysis of the LCTA data provides a large body of useful information. However, it cannot provide: (1) predicted ecological responses to natural and anthropogenic stressors or (2) the "turn key" decision-making tool required by trainers and managers. The data provided by LCTA are descriptive, and statistical inference can be made. The data are very useful in documenting what is there and what happens to it over time. But to meet the objective of supplying the Army with a training and environmental decision-making tool, we need more than descriptive data and statistical inference. We need a method for predicting ecological responses before they occur. This requires a mechanistic, rather than statistical or descriptive, model. Such a model, if adequately developed, would also become a central component of the decision-making tool used to translate ecological responses into management objectives and restrictions.
Therefore, there are two parts to this modeling effort. One is to produce a simulation model that can adequately predict ecological responses to various stressors at site, landscape, and installation levels. This model must supply site-specific information at an acceptable level of accuracy, but also must be robust enough to be used at any Army installation with a minimum amount of calibration. The second part of the task is to combine this ecological dynamics model with a decision-making model so the resulting alternative management decisions can be predicted and evaluated.
To date, we have concentrated on the first task, the development of the ecological dynamics model, because the first step is required before the second can be taken. We present an overview of its structure and an example of some of its results in the following sections. We have completed the conceptual design of the decision-making module and have started development of its software.
A primary requirement of the model is that it be able to model ecological dynamics on a mechanistic basis. The approach is to model how ecosystems function and what they consist of at some starting point. The model then produces the patterns that we recognize as disturbance and succession. The measure of success then becomes how well the predicted patterns match the actual patterns.
The current version of the model consists of four modules (climate, soil, plant, and animal), five stressors (water, nitrogen, fire, herbivory, and trampling), and three ecological processes (decomposition/mineralization, succession, and competition). Spatial and erosion components are being added to the next version of the model, and contaminants can be added now, if desired (Figure 2).
The climate module includes precipitation, season, evaporation, and light; temperature will be added at a later date. For now, the seasonal variable is adequate to account for temperature responses. The soil module divides a site-specific profile into a series of layers, each layer corresponding to a soil horizon or sub-horizon (Figure 3). Each layer has a characteristic available water holding capacity (WHC) and initial available nitrogen (N), total N, and organic matter (OM) level. The plant module, which can consider multiple plant species per community, consists of structural characteristics (including root architecture), potential growth rate, seasonal growth rate (including flowering), growth allocation factor, and water- and N-use efficiencies for each major species in the community. The animal module currently consists only of a herbivory factor. Number, biomass, and growth rate for each major animal species, along with the respective food webs, are being added.
Figure 2. Modules and variables in, or being added to, the community dynamics simulation model.

Figure 3. Schematic illustrating the structural organization of the soil and plant modules of the community dynamics simulation model.
Each variable in the model is initially calibrated based on the best site-specific data available. Precipitation is based on historical data (daily), but can also be modeled stochastically. Daily rainfall enters the profile and is distributed downward among the layers based on WHC (Figure 3). Daily evaporation can remove soil water from the upper layer. Plant roots remove soil water on a daily basis, based on potential growth rate, water-use efficiency, season, and competition with other species. Growth is also constrained by N availability and presence of contaminants. Decomposition, by layer, supplies the available N pool and is dependent on available water and organic matter.
Plants produce new growth based on maximum potential growth rate (adjusted monthly), amount of aboveground biomass (or seed biomass in the seedbank), and availability of water and N. New growth (biomass) is allocated to the respective plant parts (roots, trunk, stems, leaves, flowers/seeds) based on species-specific allocation factors. Aboveground biomass can also be removed by herbivory, fire, and senescence. Aboveground and belowground biomass can be removed by moisture stress and maintenance respiration.
Belowground competition is modeled on the basis of proportional root biomass within a soil layer and absorption efficiency of roots (species-specific). Aboveground competition is modeled on the basis of effect of shading. Once the spatial component is added, available surface space will also influence competition.
In the model, fire affects plants by removal of aboveground biomass (total in the case of a crown fire, partial in the case of a surface fire) and nutrient release. Plant response following fire is species specific (e.g., resprout, regrowth only from seeds). The fire regime can be set by month, annual frequency, and intensity. Herbivory is modeled as monthly removal of designated plant parts (e.g., leaves, leaves and stems). The herbivory rate can be altered. Trampling is modeled in a manner similar to herbivory and fire (amount and type of tissue removed).
The model provides a series of printouts giving monthly plant biomass and production (by species and by plant part), N dynamics (by soil layer, plant species, and plant part), and water dynamics (by layer). These printouts can be easily loaded into a spreadsheet for graphical displays. The spatially-explicit version of the model will display spatial patterns of species, communities, and soil variables at multiple spatial scales during the simulation runs.
Calibration of the model is currently based on LCTA and literature data. We have recently completed several greenhouse and field experiments that will supply data on the effects of moisture, nitrogen, and competition on growth and development of nine of the major species used in the model. As soon as these data are analyzed, the results will be incorporated into the model to increase its accuracy.
Field plots are being established at some of the installations and at three supporting sites. These plots will serve two purposes. First, baseline data taken at the beginning of the experiments will allow for more accurate calibration of the model for the respective communities. Second, they will be used as validation plots. Data collected from the plots over time will be used to test the accuracy of the model predictions in relation to actual field data.
We are in the process of adding the complete animal module to the model. This will allow modeling of animal population dynamics and will provide for a more realistic modeling of animal impacts on vegetation.
The current version of the model bases all dynamics on an average square meter, and does not represent the spatial heterogeneity at the multiple scales characteristic of real ecosystems. We are working on a major revision that incorporates spatial aspects of the communities and will allow for modeling of landscape-level dynamics. An early prototype version of the spatial component can scale ecosystem dynamics from 1x1 m to 100x100 m (Figure 4). Our approach is to represent multiple scales simultaneously. Most herbaceous species will be modeled on the 1x1 m scale. Shrubs and small-scale ecological processes (e.g., decomposition and mineralization; bunchgrass patchiness) will be modeled on a slightly larger scale (2x2 m), in which four replicates of the 1x1 m scale are represented. Trees will be modeled on a 10x10 m scale, with 25 replicates of the 4x4 m scale added. The 100x100 m scale will be the basic scale for simulating community-wide processes and ecotones. Larger scales, such as 1x1 km and 10x10 km, will be used to model landscape, training area, and installation-wide characteristics. In each case, aggregates of the smaller scales will be used to define the larger scales.
We have also begun developing the decision-making module to interface with the ecological model (Figure 5). When completed, this will provide trainers and managers with a PC-based tool they can use to translate ecological impacts into training and management allocation decisions (Figure 6).

Figure 4. Sample output from the prototype spatial module of the community dynamics simulation model.

Figure 5. Conceptual overview of the prototype decision-making module for the community dynamics simulation model.
Figure 6. Schematic illustrating the conceptual linkages among the ecological, management, and decision-making modules of the community dynamics simulation model.
In summary, we are developing a simulation model that can accurately predict
1. responses of ecological communities to disturbance and ecological stress
2. responses of disturbed communities to remediation/restoration efforts
3. maximum sustainable use under various scenarios and ecological conditions.
We believe this is possible because we
1. base the model on ecological mechanisms controlling ecosystem dynamics
2. calibrate a general core model to site-specific conditions for each community
3. test the model with field validation experiments
4. will adapt the model to sufficiently large scales to accommodate realistic training activity scales.
References
Diersing, V.E., R.B. Shaw, and D.J. Tazik, "The U.S. Army Land Condition-Trend Analysis Program," Environmental Management, 16(3):405-414, 1992.
Field Manual (FM) 100-5, Operations, Headquarters, Department of the Army (HQDA), 14 June 1993.
Goran, W.D., L.L. Radke, and W.D. Severinghaus, An Overview of the Ecological Effects of Tracked Vehicles on Major U.S. Army Installations, 75 pp, USACERL Technical Report N-142/ADA126694 February 1983.
Johnson, F.L., "Effects of Tank Training Activities on Botanical Features at Fort Hood, Texas," The Southwestern Naturalist, 27(3):309-314, 1982.
Macia, T.E., "Army Reinvents Training Land Protection Operators Take the Lead." in: Proceedings of the 22nd Environmental Symposium and Exhibition of the American Defense Preparedness Association, Orlando, Florida, 1996.
McLendon, T., and B.E. Dahl, "A method for mapping vegetation utilizing multivariate statistical techniques," Journal of Range Management, 36:457-462, 1983.
Severinghaus, W.D., Effects of Tracked Vehicle Activity on Higher Vertebrate Populations at Army Installations, USACERL Technical Report N177/ ADA142653, April 1984.
Severinghaus, W.D., and M.C. Severinghaus, "Effects of Tracked Vehicle Activity on Bird Populations," Environmental Management, 6(2):163-169, 1982.
Shaw, R.B., and V.E. Diersing, "Tracked vehicle impacts on vegetation at the Pinon Canyon Maneuver Site, Colorado," Journal of Environmental Quality, 19(2):234-243, 1990.
Training Circular (TC) 25-1, Training Land, U.S. Army Training Support Center (USATSC), 31 September 1991.
Trumbull, V.L., P.C. Dubois, R. Brozka, and R. Guette, "Military Camping Impacts on the Ozark Plateau," Journal of Environmental Management, 40:329 - 339, 1994.
U.S. Army Concepts Analysis Agency, Evaluation of Land Values Study, USACAA Technical Report CAA-SR-96-5, 1996.
U.S. Department of Agriculture, National List of Scientific Plant Names, Volume 1, List of Plant Names, SCS-TP-159 (Soil Conservation Service, U.S. Department of Agriculture, Government Printing Office, Washington, DC, January 1982).
USACERL DISTRIBUTION
Chief of Engineers Fort Stewart 31314 Rock Island Arsenal
ATTN: CEHEC-IM-LH (2) ATTN: AFZP-DEV ATTN: AMSMC-EHR
ATTN: CEHEC-IM-LP (2) Fort Buchanan 00934 ATTN: SMCRI-PWB
ATTN: CECC-R ATTN: AFZK-B-EHE White Sands Missile Range
ATTN: CERD-L Fort Devens 01433 ATTN: STEWS-ES-E
ATTN: AFZD-DEM US Army Dugway Proving Ground
Army Headquarters Fort Drum 13602 ATTN: STEDP-EPO-CP
ATTN: DAIM-ED-N (5) ATTN: AFZS-EH-E US Army Yuma Proving Ground
ATTN: DAMO-TRO Fort Irwin 92310 ATTN: STEYP-ES-E
ATTN: ANOP ATTN: AFZJ-EHE-EN
ATTN: AMSTE-EQ Fort Hood 76544 Aberdeen Proving Ground
ATTN: SFIM-AEC-ECN ATTN: AFZF-DE-ENV ATTN: STEAP-SH-ER
ATTN: ATIC-CTS Fort Hunter Liggett 93928
ATTN: FKEN-EN ATTN: AFZW-HE-DE National Guard Bureau 20310
ATTN: AEAGC-TD-MO Yakima Tng Center 98901-5000 ATTN: NGB-ARE
ATTN: EAGC-TD-RMD ATTN: AFZH-Y-ENR ATTN: NGB-ARI
ATTN: APOP-TR Fort Dix 08640 ATTN: NGB-ARO-TS
ATTN: APEN-EV ATTN: ATZD-EH
Army National Guard
US Army Europe TRADOC Ft Richardson, AK 99505-5800
ATTN: AEAEN-ENVR 09114 Fort Monroe 23651 Phoenix, AZ 85008-3495
100th Support Group ATTN: ATBO-FE Sacramento,CA 95826-9101
ATTN: AETT-EN-DPW 09114 Installations: Boise, ID 83705-8095
CMTC Hohenfels 09173 Fort Lee 23801 Jackson, MS 39209
ATTN: AETTH-DPW ATTN: ATZM-PTS-T Camp Shelby, MS 39407-5500
Fort Jackson 29207 Oklahoma City, OK 73111-4389
FORSCOM ATTN: ATZJ-PWN Draper, UT 84020-1776
Fts Gillem & McPherson 30330 Fort Gordon 30905 Braggs, OK 74423
ATTN: AFOP-TE ATTN: ATZH-DIE Reading, MA 01867-1999
ATTN: AFOP-TSR Fort Benning 31905 Camp Edwards, MA 02542-5003
ATTN: AFPI-ENE ATTN: ATZB-PWN Richmond, VA 23219
Installations: Fort McClellan 36205 Columbia, SC 29201
Fort Indiantown Gap 17003 ATTN: ATZN-EM Eastover, SC 29244
ATTN: AFKA-ZQ-DE-E Fort Rucker 36362 St. Augustine, FL 32085
Fort AP Hill 22427 ATTN: ATZQ-DPW-EN Starke, FL 32091
ATTN: ANAP-PWE Fort Leonard Wood 64573 Indianapolis, IN 46241
ATTN: AFZM-FHE ATTN: ATZT-DPW-EE Lansing, MI 48913
Fort McPherson 30330 Fort Leavenworth 66027 Little Falls, MN 56345
ATTN: AFZK-EH-E ATTN: ATZL-GCE Jefferson City, MO 65101
Fort Riley 66442 Fort Bliss 79916 Camp Roberts, CA 93451
ATTN: AFZN-DE-V-N ATTN: ATZC-DOE Camp Robinson, AR 72118-2200
Fort Polk 71459 Carlisle Barracks 17013
ATTN: AFZH-DE-EN ATTN: ATZE-DPW-E US Military Academy 10996
Fort Sam Houston 78234 Fort Eustis 23604 ATTN: MAEN-EV
ATTN: AFZG-DE-EM ATTN: ATZF-PWE ATTN: DOPS
Fort Lewis 98433 Fort Chaffee 72905
ATTN: AFZH-DE-Q ATTN: ATZR-ZF US Army Research Laboratory
Fort Carson 80913 Fort Sill 73503 ATTN: AMSRL-OP-SDFE
ATTN: AFZC-ECM-NR ATTN: ATZR-B
Fort Bragg 28307 Fort Huachuca 85613 US Govt Printing Office 20401
ATTN: AFZA-PW-DW ATTN: ATZS-EHB ATTN: Rec Sec/Deposit Sec
Fort Campbell 42223 Fort Knox 40121
ATTN: AFZB-DPW-E ATTN: ATZK-PWE Defense Technical Info Ctr 22304
Fort McCoy 54656 ATTN: DTIC-FAB
ATTN: AFZR-DE-E USAMC Instal & Srvc Activity
Fort Pickett 23824 ATTN: AMXEN-M 61299 102
ATTN: AFZA-FP-E 1/97