Development and implementation of a generic pasture growth model (CLASS PGM)
Introduction
The Murray–Darling Basin (MDB) in southeastern Australia is bounded in the east by the elevated peaks and plateaus (rising to 2228 m at Mt. Kosciusko) of the Great Dividing Range. Steep to moderate slopes and ranges extend westward about 300 km, to low-relief landscapes dominated by riverine sediments. Rainfall, which is highest along the eastern boundary (>800 mm/year), falls rapidly with distance as we move towards the west to average values of <400 mm/year.
Since the MDB was settled by Europeans in the early 1800s, a range of landuses have evolved across this low-relief but variable landscape, including forestry, grazing, cropping and intensive agriculture associated with irrigation. For both dry-land and irrigated agriculture, water resources are both finite and fully committed, and are influenced to a large extent by interactions between soils, topography and land use. For planning, forecasting and research purposes, there is a need for a modelling framework with generic modelling tools that can be used to evaluate the combined impacts of these factors on small to medium-sized catchments (up to 3000 km2) within the basin.
Although simulation modelling has been used at large catchment scale within the Basin, most recently in the context of managing dryland salinity (Tuteja et al., 2003, Vaze et al., 2004a), no models realistically and adequately simulate and spatially integrate the various landuses that occur across the catchment. It is not unusual, for instance, for modellers to represent the water balance impact of trees using models/frameworks developed for pastures (Weeks et al., 2005) or crops by changing parameters relating to rooting depth and Leaf Area Index (LAI).
At locations within catchments, landuse is constrained by issues related to topography (slope, drainage networks, soils, etc) and resources (rainfall, enterprise risk and benefit:cost, etc). Thus, there is some confounding between factors restraining landuse and landuse itself. For instance, in the southern sector of the MDB, plantation forestry is generally carried out in high rainfall areas (>700 mm/year) in the east; grazing with minor cropping occurs across the steeper near-slopes where rainfall is moderate (600–700 mm/year); cropping increases on lower slope landscapes (rainfall 400–600 mm/year); while broad-scale irrigation is restricted to lands in close proximity to regulated rivers and low to zero slopes.
In order to evaluate the spatial impact of this complex mix of landuse, topography, soils and rainfall on the water balance, a range of models are needed that independently and specifically caters for each of the main landuses. These need to be compatible, and spatially integrated in a catchment-based context.
CLASS (Catchment Scale Multiple Landuse Atmosphere Soil Water and Solute Transport Model) (Tuteja et al., 2004) is one such framework. It consists of a suite of six integrated models that can be used to investigate the effects of landuse and climate variability at both the paddock or enterprise level, as well as at the scale of small to medium-sized (up to 3000 km2) catchments. These six models can be used independently or as part of the whole framework, where each of them interacts with others to simulate the whole of catchment behaviour.
CLASS PGM is one of the six models developed as part of this framework. There are a range of soil moisture and pasture growth models available that use less pertinent hydrology (e.g. bucket models for water balance accounting with limited treatment of plant growth). At the other end of the spectrum are models that are very comprehensive in their treatment of plant physiology, which make them less suited for practical catchment hydrology applications. CLASS PGM has been developed with the aim to obtain a balance between plant physiology and hydrology by simulating the most important and relevant processes from both areas accurately and in appropriate detail. PGM is developed under the catchment modelling toolkit platform (www.toolkit.net.au) which uses the latest available integrated modelling approach and so it can be used with any other model developed using this platform.
This paper outlines the concepts (agronomy and water balance) and data requirements for the Pasture Growth Model—CLASS PGM (Vaze et al., 2004c, Vaze et al., 2005). Data from a long-term field experiment (Johnston et al., 2005, Johnston and Cornish, 2005) is used for validation purposes.
Section snippets
Model concepts
PGM is a mechanistic biophysical simulation model, with a daily time step for growth calculations and user-specified sub-daily time steps for the water balance calculations using U3M-1D (Vaze et al., 2004b). The model incorporates modules for carbon assimilation, herbage production and utilisation, water dynamics, nutrient and salinity stresses and animal intake/grazing. It provides a cohesive structure for analysing the behaviour of the pasture system and the complex interactions between the
Model inputs and outputs
The climatic inputs that PGM requires is a daily climate file in ‘.csv’ format with data for minimum and maximum temperature (°C), rainfall (mm), pan evaporation (mm) and total short wave radiation (MJ/m2). All the other input parameters related to the soil profile and different pasture species being simulated are entered using the graphical user interface (GUI) as shown in Fig. 1 (Vaze et al., 2004c). The simulation period (length of simulation) depends on the length (number of days) in the
Model implementation and data used
For model developers, providing confidence that the simulated system adequately portrays interactions in the ‘real’ world, as measured in experiments is a major challenge. There are several issues. Firstly, it is unusual in experiments conducted independently of model development for all parameters that the model requires to be measured. The second issue is that experimental measurements are subject to various sources of variation or error, such as measurement and sampling error and within plot
Comparison with experimental data
The fit achieved between predicted total herbage mass and measured data for the P. aquatica pasture is shown in Fig. 3. Linear regression of these data resulted in the following relationship:The experiment was not grazed until 16 September 1993, thus the experimental pasture contained a large amount of accumulated herbage prior to this time. Towards the end of the experiment the P. aquatica pasture was grazed at lower stocking rates than the
Conclusions
CLASS PGM is a generic pasture growth model which can be used to simulate growth of composite pasture types of multiple species that may be summer or winter active, perennial or annual. The model includes carbon assimilation through photosynthesis and respiration followed by tissue growth, turnover and senescence. PGM has been developed to make sure that there is a balance between plant physiology and hydrology making sure that the most important and relevant processes from both areas are
Acknowledgements
Peter Barker, Dugald Black, Ross Williams, and Charmaine Beckett have supported this work at ex-DNR. The authors gratefully acknowledge valuable contributions of the following ex-DNR staff: Geoffrey Beale, Brian Murphy, Gregory Summerell, Michelle Miller and Yuri Ivailovski. The authors also acknowledge valuable contributions of Iain Hume, John Gallant and Joel Rahman and useful suggestions from Rodger Grayson and Francis Chiew. The authors wish to acknowledge the Department of Water and
References (20)
- et al.
A theoretical analysis of grass growth under grazing
Journal of Theoretical Biology
(1985) - et al.
Predicting the effects of landuse change on water and salt balance - a case study of a catchment affected by dryland salinity in NSW, Australia
Journal of Hydrology
(2003) - et al.
Properties of porous media affecting fluid flow, Journal of the Irrigation & Drainage Division 2 (June)
(1966) - et al.
Developing joint probability distributions of soil water retention characteristics
Water Resources Research
(1988) - Johnson, I.R., 2003. A mathematical approach to modelling pastures and crops for Australian conditions, A consultancy...
- et al.
Eragrostis curvula (Schrad.) Nees. complex pastures in southern New South Wales, Australia: impact of Eragrostis curvula, Medicago sativa and Phalaris aquatica L. pastures on soil water
Australian Journal of Experimental Agriculture
(2005) - et al.
Vegetative crop growth model incorporating leaf area expansion and senescence, and applied to grass. Plant
Cell and Environment
(1983) - et al.
The Sustainable Grazing Systems Pasture Model: description, philosophy and application to the SGS National Experiment
Australian Journal of Experimental Agriculture
(2003) - et al.
Eragrostis curvula (Schrad.) Nees. complex pastures in southern New South Wales, Australia: a comparison with Medicago sativa L. and Phalaris aquatica L. pastures under rotational grazing
Australian Journal of Experimental Agriculture
(2005) - et al.
The use of a model to optimise the interaction between the frequency and severity of intermittent defoliation and to provide a fundamental comparison of the continuous and intermittent defoliation of grass
Grass and Forage Science
(1988)
Cited by (10)
Modelling vegetation dynamics in managed grasslands: Responses to drivers depend on species richness
2018, Ecological ModellingCitation Excerpt :An attempt (Lauenroth et al., 1993) to couple individual-based gap models describing changes in species composition (STEPPE) and ecosystem models describing nutrient cycling and biomass production (CENTURY) or water balance (SOILWAT) led to the development of the ECOTONE model, able to simulate transitions between biomes (Peters, 2002). Several biophysical pasture models can be used to simulate growth of composite pasture types of several species under various management strategies, such as CLASS-PGM, a generic pasture growth model (Vaze et al., 2009), or DairyMod and the SGS Pasture Model (Johnson, 2016), but these models are rather complex, limited to five “species” and not optimized for mowing management. Gemini (Soussana et al., 2012) is another very detailed biophysical model of grassland ecosystems that explicitly simulates competition among any number of clonal populations for two key resources (light and nitrogen) along vertical canopy and soil profiles; plant functional traits are used to explain the response of aboveground net primary production and of plant community structure to various soil, climate and management conditions, including grazing, cutting and fertilization.
The challenges - and some solutions - to process-based modelling of grazed agricultural systems
2014, Environmental Modelling and SoftwareCitation Excerpt :APSIM has an international user base while most of the others have a national or regional focus. Some of the notable models not included above include: the DairyNZ Whole Farm Model (Beukes et al., 2008) which is largely constructed by drawing together elements of existing models; PaSim (Graux et al., 2011), PROGRASS (Lazzarotto et al., 2009) and CLASS PGM (Vaze et al., 2009) which operate at the paddock level; and GPFARM (Ascough et al., 2010; Qi et al., 2012) which focuses on the dryland range and cropping systems of the USA. Readers are also referred to a recent review by Del Prado et al. (2013) focussing on models for greenhouse gas issues in pastoral systems.
Improving runoff estimates using remote sensing vegetation data for bushfire impacted catchments
2013, Agricultural and Forest MeteorologyCitation Excerpt :Amongst variety of hydrological models, rainfall-runoff models are widely used for surface runoff prediction due to their simple structure and few parameters (Zhang et al., 2008; Vaze et al., 2004, 2011a; Tuteja et al., 2007; Li et al., 2012a). However, lack of surface vegetation information in rainfall-runoff models may constrain their ability to predict runoff in forested catchments where actual evapotranspiration significantly affects the surface energy balance processes and water storage availability in the root-zone affects catchment runoff (Falkenmark, 1997; Yildiz and Barros, 2007; Vaze et al., 2009). This constraint makes it harder for rainfall-runoff models to accurately simulate runoff time series in the catchments where significant land surface changes occur such as wildfire.
An integrated modelling framework for regulated river systems
2013, Environmental Modelling and SoftwareCitation Excerpt :These strategies include roof and stormwater harvesting, grey water recycling and waste water reuse. Irrigation demand is the volume of water ordered by irrigators, which is governed by both crop water use factors (Vaze et al., 2009) and farmers' perceived crop water requirements. In the Source IMS, irrigation demand can be modelled using one of the two Australian crop water models, IQQM's CropMod2 (Podger, 2004; Simons et al., 1996) and PRIDE (Erlanger et al., 1992; SKM, 2007), both of which are built into the Source IMS.
Interpretive review of conceptual frameworks and research models that inform Australia's agricultural vulnerability to climate change
2011, Environmental Modelling and SoftwareCitation Excerpt :These are: Biophysical productivity models: Studies that project changes in the physical productivity of plant and animal growth, yields and physical production, often using agro-ecological models (e.g. Crimp et al., 2002, 2008a; McKeon et al., 2008; Vaze et al., 2009; Laughlin et al., 2007). Bio-economic models: Translate projected future changes in physical productivity, into projected changes in economic productivity, farm incomes, land use and regional economic impacts (e.g. Kokic et al., 2007, 2005; Nelson et al., 2007b; Heyhoe et al., 2007; Adamson et al., 2007; Crimp et al., 2008b; Schrobback et al., 2008).
Modelling water dynamics with DNDC and DAISY in a soil of the North China Plain: A comparative study
2010, Environmental Modelling and SoftwareCitation Excerpt :Data from the above-mentioned field experiment were available and were used for this study. Concerning the simulation of hydrological processes, soil water status at various temporal and spatial scales (Ranatunga et al., 2008) and water balance computations (Vaze et al., 2009) are of general importance. In addition, studies of detailed hydrological aspects are widely distributed in the literature, for instance, water transfer through the unsaturated zone (Gandolfi et al., 2006), infiltration in the vadose zone (Fahs et al., 2009), and water flow in structurally rigid porous media (Wesseling et al., 2009).