Publications

CSIRO

intro | copyright | queries | user guide | publications | bug info | data program code | feedback | search | tour | around the world | workshops

CRCCH

Abstracts for papers on Dynamic Simulations

BEVERLY, C.R. (1992): Background notes on the CSIRO Topog model. 1. Details of the numerical solution of the Richards equation in Topog_Yield. CSIRO Division of Water Resources, Tech. Memo. 92/12, 51 pp.

This report discusses and evaluates the soil moisture redistribution sub-module of the CSIRO Topog_Yield catchment model. Attention is focused on evaluating the accuracy and mass balance control of the numerical solution of the Richards equation. Speed performance of the sub-module is also assessed.

VERTESSY, R.A., HATTON, T.J., O'SHAUGHNESSY, P.J. and JAYASURIYA, M.D.A. (1993): Predicting water yield from a mountain ash forest catchment using a terrain analysis based catchment model. J. Hydrology, 150: 665-700.

The structure, capabilities and performance of a distributed parameter hydrologic model are described. The model, called Topog_Yield, permits a transient analysis of unsaturated-saturated flow and evapotranspiration to be performed across complex terrain using a one-dimensional framework. It is applied to a 0.32km² mountain ash (Eucalyptus regnans) forest catchment in the central Victorian highlands, Australia. We compare observed and predicted daily runoff values for the site over a continuous 12 year period (1972-1983) when the catchment vegetation was in an undisturbed climax condition. All input parameter values were based on published or measured data, although some variables were adjusted within the range of known variability to yield a best fit between predicted and observed stream flow in the first year of simulation, 1972. Although the model was 'calibrated' for the first year, all variables other than climatic inputs remained fixed for the following 11 years. Modelled and observed daily runoff values compare well throughout the period of simulation, despite a wide range of climatic conditions When modelled daily runoff values were lumped on a monthly basis, the model was able to explain 87% of the variation in observed monthly stream flows over the 12 year period. Modelled annual runoff was within +-5% of observed values for 6 of the 12 years of record. Annual runoff prediction errors exceeded +-10% of observed values in only 2 of the 12 years. By the end of the 12 year simulation, the model had over-predicted runoff by less than 5%. Input data requirements and model results are discussed in the light of a preliminary sensitivity analysis.

HATTON, T.J., WALKER, J., DAWES, W. and DUNIN, F.X. (1992): Simulations of hydroecological responses to elevated CO2 at the catchment scale. Aust. J. Bot., 40: 679-696.

A spatially explicit hydroecological landscape model of water, carbon and energy balances (Topog_IRM) is described. The landscape is envisaged as a catchment forested with a single stratum comprising Eucalyptus maculata trees. The model was used to simulate the direct effects of a 2x elevation in atmospheric carbon dioxide at two levels of nitrogen on catchment water yield, soil moisture status and tree growth. Experimental results used to parameterise the model are detailed. Key features of the model are (1) an ability to scale hydrological processes at the catchment scale in three dimensions, and (2) a means to integrate multiple factors/stresses on plant growth.

The effects of CO2 on catchment hydrology (water yield or soil moisture content) and forest growth (expressed as leaf area index, LAI) were modelled for a 2-year period, and contrasted with the effects of added nitrogen. Results were expressed as totals for the catchment or spatially distributed across the catchment. For the total catchment, water yield increased in the order: high CO2 low N, high CO2 with high N, ambient CO2 with low N, ambient CO2 with high N. LAI increased from 3.3 to 5.7 in the order: ambient CO2 with low N, ambient CO2 with high N, high CO2 with low N, high CO2 with high N. These results agree with previous data. New findings are:

1. with elevated CO2 a new equilibrium in transpiration is established in which leaf area increases offset decreases in stomatal conductance;
2. the addition of nitrogen increases transpiration without any indication of a new equilibrium being reached during the simulated period;
3. the spatial distribution of soil moisture changes, presenting a new resource base for spatial changes to species composition and growth rates.

The major hydroecological responses to elevated CO2 are seen as increased maximum upper canopy leaf area, increased litter inputs, especially at times of drought (hence changed fire regimes), changes in the composition of the understorey (hence litter composition, soil microfauna, and the spatial expression of biological diversity) and a slight increase in water yield.

DAWES, W. and HATTON, T.J. (1993): TOPOG_IRM. 1. Model Description. CSIRO Division of Water Resources, Tech. Memo. 93/5, 33 pp.

A computer model which simulates the water balance and plant growth across three-dimensional catchments is presented. The model is a distributed parameter, dynamic eco-hydrological model called Topog_IRM. By combining physical and physiological approaches, it is well suited to investigations into landscape response to changes in the physical and biological character of the system. Model structure and algorithms are described, including model assumptions and constraints. Data requirements are specified, including options for different soil models and the generation of climate data.

HATTON, T.J. and DAWES, W. (1993): TOPOG_IRM. 2. Example simulations. CSIRO Division of Water Resources, Tech. Memo. 93/6.

A series of simulations of soil water balance and plant growth using the hydroecological model Topog_IRM are presented. The purpose of these simulations are to show the nature of the model: the kind of input data required, the range of physical and physiological processes accounted for, and the interactions between these processes. The simulations provide an impression of the sensitivities of the model to changes in the physiological properties of plants (the factors which affect the growth of wheat, eucalyptus trees, annual and perennial pastures), soil (hydraulic properties, depth, salinity, fertility) and climate (rainfall amount and distribution).

DAWES, W.R. and SHORT, D.L. (1993): The efficient numerical solution of differential equations for coupled water and solute dynamics: the Waves model. CSIRO Division of Water Resources, Tech. Memo. 93/18.

Systems of equations are described that allow simultaneous modelling of the dynamics of moisture and dissolved solute in a vertical column. These equations are solved with computational speed and reliability normally associated with simplified models of water balance only. The solute is assumed not to be affected by absorption/desorption or by chemical reaction, is non-volatile, and does not affect hydraulic properties. Mass balance of both water and solute equations is exact; only numerical round-off error need occur. Solute equations are linear, and modelling requires much less computational effort than water dynamics. Implications of non-conservative solutes are described.

Application of these solutions is shown for a flood plain model, called WAVES. The model includes a surface energy balance, surface carbon balance (for vegetation growth), periodic flooding at the surface, a slowly changing groundwater table at the lower boundary, and full moisture and solute dynamics.

HATTON, T.J., DAWES, W.R. and VERTESSY, R.A. (1995): The importance of landscape position in scaling SVAT models to catchment scale hydroecological prediction. In: R.A. Feddes (ed.) Space and Time Scale Variability and Interdependencies in Hydrological Processes, pp. 43-53 (Cambridge University Press: Cambridge).

Process-based models of soil-vegetation-atmosphere interactions developed for small plots (points) define vertical transfers of water and energy. One can attempt to scale to larger heterogeneous land units by disaggregating the landscape into a set of elements and applying a vertical SVAT model independently to each element (Running et al., 1989; Pierce et al., 1992). Such applications fail to consider lateral transfers. A distributed parameter, three-dimensional SVAT (Topog_IRM) developed by the CSIRO Division of Water Resources (O'Loughlin, 1990; Hatton and Dawes, 1991; Hatton et al., 1992) is used to examine the importance of lateral transfers of water for prediction of water balance components at the small catchment scale.

Simulations are used to contrast the predicted water balances from a SVAT model with and without considerations of lateral subsurface and overland flow in complex terrain. Components of the catchment water balance are shown to scale linearly except in those cases where transient perched water tables develop in landscapes with sufficient slope and hydraulic conductivity to redistribute water effectively via subsurface lateral flow. In such cases, the prediction of catchment yield and the spatial pattern of soil moisture requires the explicit treatment of lateral transfers.

HATTON, T., DYCE, P., ZHANG, L. And DAWES, W. (1995): Waves - An ecohydrological model of the surface energy and water balance: sensitivity analysis. CSIRO Division of Water Resources, Tech. Memo. 95/2.

A sensitivity analysis of WAVES, a complex biophysical process-based model that simulates movement of water, energy, and solutes in a vertical soil-plant-atmosphere system, was performed on a data set from a small cropping catchment near Wagga Wagga, New South Wales. The model was first calibrated to closely reproduce the soil water balance and plant growth of the wheat crop. A set of 15 key model parameters, representing the soil and plant variables most likely to control model behaviour, were tested by perturbing their base values by +-10%. It was shown that, of the plant parameters, the maximum assimilation rate of CO2 was by far the most sensitive with respect to changes in predicted LAI, transpiration, and deep drainage. Of the soil parameters, the saturated hydraulic conductivity had the greatest influence on deep drainage, followed by the sorptivity (lambda) and shape (C) soil hydraulic parameters of the Broadbridge and White (1988) soil hydraulic model. The paper also describes the WAVES model formulation to aid sensitivity interpretation.

VERTESSY, R.A., HATTON, T.J., BENYON, R.J. and DAWES, W.R. (1996): Long term growth and water balance predictions for a mountain ash (E. regnans) forest subject to clearfelling and regeneration. Tree Physiology 16:221-232.

We used a physically based ecohydeological model to predict the water balance and growth responses of a mountain ash (Eucalyptus regnans F. Muell.) forest catchment to clear-felling and regeneration. The model, Topog_IRM, was applied to a 0 .53 km² catchment for a 3-year pretreatment period, and a 20-year period following clear-felling and reseeding of 78% of the catchment area. Simulations were evaluated by comparing observed and predicted streamflows, rainfall interception and soil water values. The model faithfully simulated observed temporal patterns of overstory live stem carbon gain and produced a leaf area trajectory consistent with field observations. Cumulative throughfall was predicted within 1% of observations over an 18-year period. Over a 4-year period, predicted soil water storage in the upper 1.5m of soil agreed well with field observations. There was fair correspondence between observed and predicted daily streamflows, and the model explained 76% of the variation in monthly flows. Over the 23-year simulation period, the model overpredicted cumulative streamflow by 6%. We argue that there is a useful role for physically based ecohydrological models in the management of mountain ash forest catchments that cannot be satisfied by simple empirical approaches.

ZHANG, L., DAWES, W.R. and HATTON, T.J. (in press): Modelling hydrologic processes using a biophysically based model - Application of Waves to FIFE and HAPEX-MOBILHY. In press, J. Hydrology.

A biophysically based model (WAVES) is described which predicts the dynamic interactions within the soil-vegetation-atmosphere system. The physiological control on transpiration is a canopy resistance calculated as a function of net CO2 assimilation rate, the vapour pressure deficit and the CO2 concentration at the leaf surface. The soil hydrology is modelled using the Richards equation and evapotranspiration is calculated using the Penman-Moteith equation. A unique feature of the model is the modification of the vapour pressure deficit below the canopy as a function of the degree of atmospheric coupling.

The model was tested using the energy flux and soil moisture measurements from the First ISLSCP Field Experiment (FIFE) and Hydrologic Atmospheric Pilot Experiment and Modélisation du Bilan Hydrique (HAPEX-MOBILHY). The simulated net radiation, evapotranspiration and soil moisture content agreed well with the observation and with previous studies of transpiration and soil evaporation. The success of the model was due to the reasonably realistic treatment of the soil canopy processes. The utility and limitations of the model are discussed.

DAWES, W.R., ZHANG, L., HATTON, T.J., REECE, P.H., BEALE, G. And PACKER, I. (in press): Application of a distributed parameter ecohydrological model (Topog_IRM) to a small cropping rotation catchment. In press, J. Hydrology.

A biophysically-based distributed parameter catchment model, TOPOG_IRM, is described which predicts the dynamic interactions between soil-vegetation-atmosphere systems. The physiological control on transpiration is formulated using the canopy resistance defined as a function of the net assimilation rate, the relative humidity and CO2 concentration at the leaf surface. Rainfall infiltration, runoff and redistribution is simulated with the Richards equation and evapotranspiration is calculated using the Penman-Monteith equation. Two unique features of the model are (i) to couple the system by changing the value of the saturation vapor pressure deficit of air in the canopy, and (ii) the plant carbon balance which allows the simulation of plant growth using a saturation rate kinetics formulation. The model was validated using evapotranspiration, soil moisture and leaf area index measurements from Wagga Wagga, NSW in Australia for a period of 1992 to 1993. The calculated evapotranspiration was in a good agreement with the observations. The soil moisture content at various depths was well simulated for two typical sites. The model also reasonably reproduced leaf area index of wheat and canola for two growing seasons. The success of the model simulations was due to the reasonably realistic treatment of the soil and canopy processes. The sensitivity analysis indicated that (i) the maximum assimilation rate if carbon affects canopy transpiration significantly, and (ii) the total drainage is sensitive to the lower boundary conditions and the saturated hydraulic conductivity. TOPOG_IRM is a valuable tool in studying catchment responses under different land management practices. However, the application of the model is limited by the large amount of data regarding soil and vegetation properties, and their spatial distribution.

VERTESSY, R.A. DAWES, W.R., ZHANG, L., HATTON, T.J. and WALKER, J. (1996) Catchment scale hydrologic modelling to assess the water and salt balance behaviour of eucalypt plantations. CSIRO Division of Water Resources, Technical Memorandum 96/1, 23 pp.

The establishment of broad scale plantations is being considered in extensive areas of Australia as a means of remediating land degradation problems, brought about by the clearing of native vegetation for agricultural purposes. A particular focus is their use in managing waterlogging and salinity problems. However, little is known about the long term ability of plantations to modify water and salt balances at the catchment and regional scales. In this paper we summarise our experience in the development and application of a model, Topog_Dynamic, which is well suited to predicting the carbon, water and salt balance of plantations. The model can be used to determine which parts of the landscape are best suited to plantation establishment and help managers to define what the likely environmental benefits of plantation establishment are in terms of waterlogging and salinity remediation.

WU, H., RYKIEL Jr.,E.J., HATTON, T. and WALKER, J. (1994) An integrated rate methodology (IRM) for multi-factor growth rate modelling. Ecol. Modelling 73: 97-116.

The integrated rate methodology (IRM) is a method for calculating a plant growth index to estimate the effective relative growth rate from a potential relative growth rate. IRM is a procedure for constructing a single growth multplier equation that integrates multiple environmental factors affecting plant growth into a number between zero and one. The method can be used to simulate growth dynamically. The IRM framework can be used for both qualitative and quantitative modelling where the growth rate of a given entity or quantity is relevant. There are many applications in biological, economic and engineering areas. Here we develop the methodology in the context of plant growth modelling. We show how to treat multiple factors and how to include interactions. The methodology strikes a compromise between simplicity and complexity for the problem and data at hand, and can easily be used as a component of larger models.

SILBERSTEIN, R.P., VERTESSY, R.A., MORRIS, J. AND FEIKMA, P.M. (1999):
Modelling the effects of soil moisture and solute conditions on long term tree growth and water use: a case study from the Shepparton irrigation area, Australia. Agricultural Water Management 39:283-315

This paper discusses the growth and hydrologic impact of a small (2 ha) 21-year old plantation growing over a shallow saline water table at Kyabram, in the Shepparton irrigation area. Using TOPOG_Dynamic, an ecohydrological model, we simulate leaf and stem growth, water use and salt accumulation of the plantation, and assess likely future trends. The model results are compared to measurements of leaf area index, transpiration, and water table dynamics over a 4-year period, and to water table dynamics, stem growth, and salt accumulation over the life of the plantation. The model generally performed well but tended to underestimate stem growth, and bias root growth too heavily towards the surface. Shallow root activity affects the depth of the majority of salt accumulation within the soil column. Simulations are used to examine the effects of water table depth at planting, groundwater salinity and tree salt sensitivity on growth, salt accumulation and water table dynamics. The effect of depth to water table at the time of planting is found to be highly dependent on the salinity of that water table. If the groundwater salinity is 700 mg per litre the water table is drawn down, but if the water has a salinity of 6000 mg per litre, then after an initial period of drawdown, the water table rises again. The sensitivity of the plantation to salt concentration appears to have the most significant influence on growth, especially if groundwater has a salinity above 2000 mg per litre. Simulations of harvesting a plantation and returning the site to pasture suggested that there may be a severe degradation of future pasture production on that site, as the salt which had accumulated beneath the plantation rises into the root zone of the pasture when irrigation recommences. It is likely that a substantially increased leaching fraction would be required to negate this salt rise. When the plantation is irrigated, wood production increases, but the increased water requirements, and lesser environmental benefit from a shallower water table would need to be considered before irrigation could be recommended. The simulations presented here suggest that finding the best plantation rotation involves a balance between pasture and plantation productivity.

DAVIS, S.H., VERTESSY, R.A. AND SILBERSTEIN, R.P. (1999):
The sensitivity of a catchment model to soil hydraulic peoperties obtained by using different measurement techniques. Hydrol. Process. 13: 677-688

Most studies on the use of physically based hydrological models have identified saturated hydraulic conductivity (K_sat) as one of the most sensitive input parameters. However, K_sat is also one of the most difficult landscape properties to measure accurately, casting doubt on the ability of modellers to estimate this parameter a priori for catchment simulations. Several studies have shown that K_sat estimates are greatly influenced by the measurement method used, primarily because of scale effects. In this paper, we evaluate the effect of K_sat measurement method on catchment simulations aimed at predicting water yield from forested catchments. A series of simulations are conducted using the TOPOG_Dynamic catchment model, with K_sat estimated by means of the constant head well permeameter, small core (6-3 cm x 7.3 cm) and large core (22.3 cm x 30 cm) methods. These were applied in a deep, permeable forest soil in which macropore flow has been noted to occur. The three measurement methods yielded very different K_sat estimates and these had a large effect on model results. The model predictions based on small core and well permeameter measurements were extremely poor, as these methods did not adequately account for preferential flow through the soil. The large core estimates of K_sat, which were one to three orders of magnitude higher than the values obtained by the other two techniques, produced good predictions of catchment discharge and known spatial patterns of water table depth. Our results highlight the need for caution when applying soil hydraulic measurements to catchment-scale models.

ZHANG, L., DAWES, W.R., HATTON, T.J. HUME, I.H., O'CONNELL, M.G., MITCHELL, D.C., MILTHORP, P.L. AND YEE, M. (1999):
Estimating episodic recharge under different crop/pasture rotations in the Mallee region. Part 2. Recharge control by agronomic practices. Agricultural Water Management 42:237-249

Much environmental degradation, including salinity in the Mallee region of southeastern Australia, is associated with the loss of native vegetation and increased recharge. As a result, various agronomic practices have been proposed to reduce groundwater recharge. This study was conducted to evaluate the impact of these practices on recharge, in particular episodic recharge. A biophysically based model (WAVES) was used to estimate recharge rates under some typical crop and pasture rotations in the region using long-term meteorological data. Results show that: (1) recharge just below the root zone was episodic and that just 10% of annual recharge events contributed over 85% of long-term totals. Management options such as incorporating luceme and deep-rooted non-fallow rotations can reduce both, mean annual recharge, and the number of episodic events, but not eliminate recharge completely; (2) winter fallows increased soil-water storage and some of the additional water was stored in the lower portion of the root zone or below it. This can increase the risk of recharge to groundwater system; (3) changes in land management may take a considerable period of time (>10 years) to have any noticeable impacts on recharge; and (4) recharge under lucerne was around 30% of that under medic pasture.

SILBERSTEIN, R.P. AND VERTESSY, R.A.
TOPOG_Scenario - A tool for exploring hydrologic possibilities. 2nd Inter-Regional Conference on Environment-Water 99

Hydrological models are used both as a means of testing hypotheses on, and applying results from, data collected from detailed and often expensive field experiments. The process of experiment and modelling is an iterative one, in which ideas and understanding of processes can be tested against the data and modified when necessary. TOPOG_Scenario is a new package which enables us to set up and run hundreds computer simulations of catchments in a single organised batch, and assemble the output in a manageable form. We describe how we preselect a range of soils, climates, catchments, and vegetation distributions for simulation, and select output variables for analysis. After each model run, the selected output is analysed and summarised into a small number of composite files which contain the output of all simulations. The output files can be input directly to a graphics package and response surfaces of the model parameter space plotted. For large batches, with a large number of permutations, this can be facilitated with the use of a database package to aid the query design and extraction of suitable output results. The package has been used to study possibilities for using trees in agroforestry to control water balance problems and preserve catchment productivity, the essential results of which are presented in a companion paper.

Sobieraj, J.A., Elsenbeer, H. and Vertessy, R.A.
Pedotransfer functions for estimating saturated hydraulic conductivity: implications for modeling storm flow generation. J. Hydrology 251: 202-220

We evaluated the performance of nine published pedotransfer functions (PTFs) for estimating saturated hydraulic conductivity (K_s) in modeling the stormflow generated in rainforest catchment. Using available input data consisting of particle size distribution, bulk density, and saturated moisture content information, these empirically-based PTFs were found to be inadequate in estimating K_s in this catchment. At shallow depths (0-0.1 m), PTFs commonly underestimated K_s by variable amounts with the exception of the Jabro PTF, which either overestimated K_s or was not significantly different from measured values. At subsequent depths (0.1 - 0.4 m), PTFs typically overestimated K_s by variable amounts, the exception being the Campbell and Shiozawa PTF, which typically underestimated K_s . We used TOPOG_SBM to model storm flow generation by replacing measured K_s values from 0 to 0.1 m depth with PTF-estimated K_s values. The simulation set using Rosetta SSC (PTF with input of % sand, silt, clay) K_s values overestimated runoff for all events, and overland flow occurred across the entire catchment for all events. Simulations using Rosetta SSC-BD (PTF with input of % sand, silt, clay and bulk density) K_s values predicted hydrograph attributes as well as simulations using measured K_s values, but the Rosetta SSC-BD simulation set predicted a much larger spatial frequency of overland flow across the catchment that the measured K_s simulation set. Model simulations using the Jabro PTF, which generated large estimations of K_s , produced hydrographs that overestimated total runoff and time of rise, but underestimated peak runoff. This model predicted much less overland flow than other models. currently published PTFs used in this study are inadequate in estimating K_s for the La Cuenca catchment, which in turn make them inadequate for modeling storm flow generation. enhanced model performance could likely be achieved by utilizing PTFs that better account for the influence of macroporosity.

---