ModMED logo ModMED Home page -  Aims and Objectives - Software -  Participants -  Publications - 
The following is a short paper (currently in press) that summarises the aims and objectives and modelling approach of the ModMED Project. It is based on a poster presented at the International Conference on Mediterranean Desertification, Crete, in October 1996.

ModMED
Modelling Mediterranean Ecosystem Dynamics

Stefano Mazzoleni: Istituto di Botanica, Universitá di Napoli, I80055 Portici, Italy
Colin Legg, Robert Muetzelfeldt: Institute of Ecology and Resource Management, The University of Edinburgh, King's Buildings, Mayfield Road, Edinburgh EH9 3JU, UK.
Abstract
Introduction
The Approach
Data collection
The Modelling Environment
Results and Discussion
References

Abstract

Recent developments in agricultural policy and land use have resulted in major changes in the ecology of Mediterranean ecosystems. Changes in grazing patterns and fire frequency and intensity have been major contributing factors. The ModMED project is concerned with the dynamics of vegetation in response to these changes. The relevant ecological processes at the landscape, community and individual levels are being researched in Italy, Portugal and Greece and represented in a hierarchical, modular modelling environment. The ModMED modelling environment provides a powerful tool for scientific research into the ecology and management of semi-natural vegetation, and for decision making in resource management at regional and national levels.

Introduction

Mediterranean ecosystems are currently under threat from several types of disturbance leading to desertification and a reduction of the productivity of agricultural land for human use, as well as the possible loss of biodiversity and wildlife habitats. To predict the response of an ecosystem to disturbance, it is necessary to understand the processes determining vegetation community structure and functions. The ModMED Project is modelling the response of Mediterranean shrub-dominated vegetation to disturbance by fire and grazing to facilitate appropriate resource management decisions.

Vegetation comprises many interacting individuals of many plant species. The competitive balance between species, and hence the successional development of vegetation, is largely determined by patterns of grazing and fire frequency. The maquis and garrigue vegetation has evolved over thousands of years in an environment of heavy grazing and frequent cutting and burning but agricultural land use in the Mediterranean has changed considerably in the last 20 - 30 years. Records show that grazing of semi-natural Mediterranean vegetation by sheep and goats has virtually ceased in several European countries where marginal land has been abandoned. This has lead to the rapid succession to woodland and accumulation of biomass. This in turn then affects the frequency and intensity of fires. Abandonment of traditional land management will result in dramatic changes to the landscape.

Vegetation plays a vital role in the protection of soil structure through its effects on hydrology and therefore in preventing desertification. ModMED will provide a tool to predict changes in the structure and function of vegetation in response to changes in fire and grazing regimes.

The Approach

The primary objective of the ModMED project is to predict the development of vegetation patterns in the landscape in response to changes in land use. It is an underlying assumption that successful prediction of changes in the landscape will come from an understanding of the processes of succession and vegetation change in response to disturbance. Similarly, prediction of community development will come from an understanding of the behaviour of individual plants. Research is therefore being conducted on factors relevant to vegetation dynamics at a wide range of spatial and temporal scales. Data and information on ecological processes are being collected from the literature and through field survey and experiments in Italy, Greece and Portugal.

Data collection

At the landscape level, ground survey and aerial photographs taken from the 1940s onwards are being used to map the past and present distribution of vegetation and soil types in a GIS. Old photographs and documents give evidence of former land use (Di Martino, 1993). Knowledge of the history of management is essential to understanding the present structure and distribution of these anthropogenic vegetation types.

At the community level, research is concerned with post-fire succession and the influence of grazing on vegetation. Regular monitoring of permanent plots provides estimates of rates of change in vegetation of different ages. Measures of fuel quantity and quality permit the prediction of fire behaviour, while the behaviour of grazing animals is related to the distribution of fodder. Litter production and its role in soil development provides a link between vegetation and hydrology. Studies of seedling demography following fire complement manipulative experiments on factors affecting seedling establishment.

At the individual level, measurements of the water balance and photosynthesis of several species growing in their natural environment show the physiological tolerances and responses of individual plants to their immediate microenvironment. This information relates to the phenology and growth of plants and their response to damage by grazing animals.

The Modelling Environment

A modular modelling approach enables the user to construct a model using a library of compatible modules (Muetzelfeldt, et al., 1996; Donatelli, 1995; McCown et al., 1996). ModMED provides a flexible modelling environment that permits the user to model Mediterranean ecosystems in a variety of ways. Each module may model a different part of the system, and there may be a choice of modules for modelling the same part in different ways. The hierarchical approach (Laval, 1995) divides the system into the landscape, community and individual levels. The landscape-level model may contain many different community-level models of different types. Each community-level model may contain representations of many different individuals. Modularity and compatibility of modules is maintained using object-oriented programming.

The Landscape-level Model comprises a raster representation of the landscape using topographic and soil data from a GIS. Processes modelled at this level include long-range seed dispersal, the distribution of grazing animals and the spread of fire (Table 1). The user controls fire initiation and stocking rates during the simulation and can monitor model predictions through maps and tabular output of data.

Table 1.
Examples of ecological processes that are modelled at each organisational level in the ModMED Landscape Model and Individual-based Community Level Model.

Landscape Level Distribution of grazing animals in the landscape
Long-distance seed dispersal
Spread of fire
Climate

Community Level Light penetration through plant canopies
Seedling establishment
Litter accumulation and soil processes
Soil moisture relations
Species selection by grazing animals
Fuel quantity and quality and fire intensity

Individual Level Water uptake and leaf water potential
Changes in stomatal resistance
Photosynthetic responses to light intentsity and water stress
Phenological patterns and responses to grazing damage
Allocation patterns to root, shoot, stem, etc.

Each 'cell' or spatial unit within the Landscape model contains a community-level model of vegetation (Figure 1). This could be a simple Markov model of transitions in the state of the vegetation. Alternatively, more complex models of vegetation dynamics may be based on qualitative rules, compartment flows or population dynamics. The most complex model being developed by ModMED includes a spatially-explicit representation of individual plants. This, the Individual-based Community-level Model, includes the processes of soil development and hydrology, fire intensity and seedling establishment. Light interception by vegetation and offtake by grazing animals are also calculated. The individual plants within the community respond to their immediate environment of light intensity and soil moisture potential with a physiologically-based model, and respond to the effects of grazing and burning through the allocation to root and shoot growth and the development of a canopy shape. The behaviour of each individual depends on the ecological attributes of the species.

Figure 1. The Hierarchical Structure of the ModMED Model.
A landscape comprises a grid of pixels containing one of several vegetation types (a and b). Each grid can be modelled in one of several ways, for example as a Markov model (e.g. probability of transition from state a to state b is 0.3), or as a spatially explicit collection of individuals. Each individual may be modelled at the physiological level.

Each spatial unit in the landscape level holds a community-level module which is represented as existing in one of a number of states (a or b). The community-level transitions between states may be modelled, for example, as a simple transition-matrix model, or as a complex spatially-explicit individual-based model. Each individual plant is modelled at the physiological level.

Results and Discussion

The ultimate objective of ModMED is to be able to predict the future pattern of vegetation in the landscape given certain environmental and management conditions. In particular, the effects of changes in grazing and fire will be modelled. This will provide a powerful tool for the management of semi-natural vegetation and for decision making in resource management at regional and national levels. For example, the model can be used for investigating the effects of abandonment of grazing on fire frequency and intensity, or for testing the potential for vegetation management by prescribed fire.

Model outputs at the landscape level can be in the form of proportions of the landscape that are occupied by different vegetation types, or as maps in the form of a GIS. At the community and individual levels the model produces graphs of changing micro-environment and physiological status of individual plants as the vegetation develops in a range of burning and grazing management regimes.

These types of management will have important consequences for soil and water conservation. The model will predict water usage of vegetation given different weather conditions. The accumulation of litter and soil organic matter also controls the water-holding properties of the soil. A hydrological module is being developed as part of the ERMES Project which will provide a link between vegetation dynamics and other related research on hydrology, climate change and desertification.

Validation is an important part of any modelling exercise and much of the data collection is for testing the model predictions at the community and landscape levels. The process of constructing, parameterising and testing a model is fundamental to the scientific method. The ModMED model will provide a valuable tool for future research as more of the ecological processes are investigated and parameterised.

The model also has valuable educational potential and a simplified version has been produced which is linked to a hypertext display for use in schools, and other educational establishments.

References

Di Martino, P. 1993. Deforestation and natural regeneration of woodland: the forest history of Molise, Italy, over the last two centuries. In: C. Watkins (ed.) Ecological effects of afforestation. CAB International, pp. 69 - 92.

Donatelli, M. 1995. CSYMBA (Cropping Systems Model Building Assistant) Program Documentation and User Manual. Version 1.0 Beta. Istituto Sperimentale Agronomico - Sezione de Modena. Italian Ministry of Agriculture, Food and Forestry Resources.

McCown, R. L., Hammer, G. L., Hargreaves, J. N. G., Holzworth, D. P. & Freebairn, D. M. 1996. APSIM: a novel software system for model development, model testing and simulation in agricultural systems research. Agricultural Systems 50: 255 - 271.

Laval, P. 1995. Hierarchical object-oriented design of a concurrent, individual-based model of a pelagic tunicate bloom. Ecological Modelling 82: 265-276.

Muetzelfeldt, R. I., Taylor, J., Wiggins, G. & Sharp, L. 1996. Modular approaches to agroforestry modelling. In Agroforestry Modelling and Research Coordination: Annual Report on ODA FRP R5652. Chapter 11. Institute of Terrestrial Ecology.


ModMED Home page -  Aims and Objectives -  Software -  Participants -  Publications
All comments on these pages to: Colin Legg
Institute of Ecology and Resource Management
The University of Edinburgh
Darwin Building, King's Buildings
Mayfield Road
Edinburgh EH9 3JU
Scotland, UK
Last update: 4/4/2000