Previous PageTable Of ContentsNext Page


Grassland resource assessment is inextricably linked to all other components of the pastoral production system and can only be carried out realistically as part of an integrated land, forage and livestock resource analysis. Analyses are generally carried out to quantify resource endowment, understand interrelationships between resource components, predict environmental impact, estimate livestock support capacity, and to appraise development options.

Resource assessments using traditional methods are often difficult and time consuming, and carry the risk of misleading results, due to a lack of accounting for the complex and interrelated nature of forage supply and livestock forage requirements - characteristics typical of most pastoral systems. The traditional simple DM balance does not account for the significant qualitative differences between forages, nor for livestock feed requirement variations due to the growth and physiological state of the animal, and nor does it account for forage carryover losses.

Modelling provides an opportunity to better understand the complexity of a pastoral system and therefore is consistent with the "new paradigm" of understanding grassland ecosystems. It enables fast repetitive evaluation of forage resource and livestock data at both component and system levels. Analyses spanning time, incorporating variability and more realistically representing the complex interactions within the system, become possible using modelling.

Several packages are available for modelling agriculture and livestock production (Box 11). Most are designed for intensive production systems and only a few are suited to livestock production on extensive grasslands and the forage production systems and analyses of the type described in this document.

Field application of some grassland modelling systems is restricted because they require a wide range of primary environmental data, such as temperature, precipitation, relative humidity, wind velocity, radiation and a range of information on soil conditions. These are not available for the majority of sites subject to grassland development projects. The use of primary environmental parameters may not account for plant community heterogeneity and the impact of historical grazing upon the current status and potential productivity of the grasslands. Furthermore, models may not account for factors such as occurrence of pests and diseases, adverse local conditions, nor account for socio-economic and political factors that influence herder and farmer decisions (Schultink and Amaral, 1987).

To be a practical tool for analysing pastoral resources of the types common in development projects, a model must meet the following criteria:

The relevance of a model to the objectives of the development programme and to the type of pastoral environment should be critically assessed prior to application, otherwise potential benefits will be lost in data compilation and set-up time, and there will be a major risk of system misrepresentation. Land, forage and livestock databases used in the models must be realistic and reflect practicable scenarios of biological function, management policy and environmental opportunity and potential.

Box 11.  Computer-based tools for modelling crops and livestock production on pastures and extensive grasslands

Several computer-based tools are available for predicting forage and crop production, modelling farms and extensive grassland-based pastoral systems, and assessing the economics of various strategies. Examples are:

- The YIELD module of the CRIES (Comprehensive Resource Inventory and Evaluation System), developed for crops (USA).
- EHRYM is a model for yield prediction based upon temperature, precipitation and solar radiation (USA) (Fisser, 1986).
- DSSAT (Decision Support for Agrotechnology Transfer), combines crop, soil and weather information to simulate multi-year crop management strategies (USA).
- RANGEPACK HerdEcon, a dynamic herd or flock model that is linked to property cash flows (Australia) (Stafford Smith and Foran, 1992).
- RAPS (Resource Assessment for Pastoral Farming Systems), a mechanistic model and decision-support system for assessing forage resources, livestock carrying capacities and development options in complex pastoral systems (New Zealand) (Harris, 1998).
- GRAZE, a mathematical model that simulates daily performance and interactions associated with beef-forage grazing systems. Accounts for prevailing weather, edaphic conditions and growth status of forage plants (USA).
- STOCKPOL models sheep, deer and cattle farms and evaluates livestock production and management options (New Zealand).
- GRAZFEED, a grazing management tool that predicts cattle and sheep production from available forage and other feed sources (Australia).

The quality of model output depends on the quality of the input data sets. Sophisticated models, including graphical presentation of results, carry the risk of user acceptance of output without critical review. Rubbish input merely results in sophisticated looking rubbish output.

RAPS (Resource Assessment for Pastoral Systems; Figure 14) is a computer-based model designed specifically for use in pastoral development programmes, and operates on IBM PCs and compatibles as a stand-alone, fully compiled programme consisting of approximately 50 000 statements. It aids managerial, policy and development decisions, primarily by analysing land and forage resources for their productivity and livestock support capacity, taking system complexity into account. The origin of the programme is in the evaluation of New Zealand pastoral properties during the late 1970s. Development of the programme began in earnest during the mid-1980s, and since then it has undergone considerable enhancement and testing. The major components of the programme have more than twelve years of field application in a wide range of environments and countries.

RAPS has been used as a pastoral resource assessment tool in environments ranging from humid mountains to arid deserts. Livestock types have included sheep, goats, cattle, horses, camels and yaks. It has been used in the assessment of private pastoral properties, in local and regional development projects and for land use planning.

Figure 14.  RAPS main screen

Figure 15.  The scope of input parameters used by RAPS and types of output

RAPS primarily, but not exclusively, uses metabolizable energy (ME) as the forage-livestock integrator for modelling a pastoral system and draws upon a wide range of data sets and variables. The scope of data used to represent and analyse pastoral systems is summarized in Figure 15 together with types of output.

The primary data sets and information required are:

For land and forage units:

And for livestock:

The forage database edit screen is shown in Figure 16.

Figure 16.  RAPS forage database edit screen

Figure 17.  Simplified flow chart of the RAPS kernel

Most data sets are represented as seasonal profiles with a resolution of 24 points per year. In the modelling process, none of the grassland types or other forage sources are aggregated and, unless management constraints or other use patterns are imposed, the forage resources are prioritized for use on the basis of livestock forage quality requirements and minimized sequential forage quality "loss." Figure 17 gives a simplified flow chart of the RAPS kernel mechanism.

Formulae for determining energy requirements for varying body weights and physiological condition are based on a number of internationally published standards (e.g. Agricultural Research Council, 1980; National Research Council, 1981; Standing Committee on Agriculture, 1990) and modified according to local conditions. Where forage quality data is not locally available, published standards for equivalent forage types are used (e.g. Cottle, 1991; Drew and Fennessy, 1980).

Because RAPS uses "high-level" input parameters for the forage and livestock components (i.e. not primary environmental factors), it can utilize a variety of data. Sources are normally: government statistics, research results, surveys, farmer and herder interviews, and field inspections. In practice, the relative contribution of each source depends upon the availability and quality of the information. Rarely is data compiled from only one or two sources.

The importance of accounting for pastoral system dynamics, especially intra- and inter-year variability in forage supply, is emphasized throughout this document. Forage resource dynamics are accounted for in the model by the inclusion of statistical coefficients of variation of forage production within each of the 24 intervals of the year. Each source of forage may have its own characteristic profile of variability. Thus irrigated pasture will have a significantly different pattern and level of variability compared to rainfed semi-arid grassland.

Another aspect of resource analysis that has been emphasized as a shortcoming of traditional forms of "carrying capacity" estimation is the carryover of standing forage from one time interval to another until it is required for grazing. The model uses coefficients of depreciation for physical forage loss and forage quality depreciation for each forage type. Within an interval, a forage type is divided into up to three strata and the carry-over processes within each stratum may be handled differently. The significance of quantitative and qualitative depreciation of forage depends on the level of "incompatibility" of the seasonal patterns of forage production and the patterns of livestock forage requirement. Carry-over losses are commonly in the range of 5% to 35%, and "losses" due to prescribed utilization limits for a particular forage type are in addition to this.

A potential constraint to the use of ME as the "common denominator" of a model is the difficulty in estimating the actual intake by (and therefore growth rates and productivity of) a particular livestock type for a particular type of forage. This problem is avoided in the model because the livestock growth rates and physiological status (e.g. pregnancy and lactation) are defined to represent a particular livestock regime. This enables a highly realistic representation of the production system. Different livestock production scenarios, for example the "current" system and an "improved system," are described in different databases. The effect of differences in forage, particularly quality, on livestock intake requirements to meet prescribed growth and production targets is handled dynamically by the model.

The first step in modelling a pastoral system is to compile and test baseline land, forage and livestock databases. These represent the current productivity, utilization levels and management practices of the pastoral system. To represent the current pastoral system it is necessary that:

  1. Forage yields are those actually occurring under present management and not those regarded as potential under optimum grazing management or forage development.
  2. Forage yield estimates are those of a "normal" year and as such may be used as a benchmark for assessing the impact of inter-year variability of yields.
  3. Forage quality indices, and associated depreciation schedules, reflect the quality of the current forage resources.
  4. Utilization levels represent actual levels, not those considered optimal.
  5. Livestock types, number, liveweights and levels of feeding (sometimes under-feeding) are those normally experienced, and not optimized.
  6. Normal seasonal patterns of grazing use of the land or forage units are applied.

The integrity of baseline databases are checked using convergent indicators to ensure that the pastoral system is realistically represented. Indicators include:

Once the baseline databases have been compiled and tested, copies are modified to represent development options and other scenarios. The potential applications and uses of computer-based pastoral system modelling is far greater than traditional methods, since modelling accounts for system complexity and component interrelationships. Examples of the use of modelling include:

Chapter 6 presents examples of the application and output of the model in four case studies, in reverse chronological order, so the Gansu study illustrates use of the most recent RAPS version.

Previous PageTop Of PageNext Page