This question is more general, of course. Why do we need modelling in hydrology, in science? What does it mean to explore nature? Is it to decompose it into many parts and describe them? Is it to observe the continuation of processes and accumulate measurement data? Does it mean to look for dependencies of the components on each other? All these are necessary but will not bring us to understanding of nature as a holistic dynamical organism. To explore nature it is necessary to generalize our knowledge and test it in real conditions that is to create a mathematical model describing natural processes. “Re-creating” nature in our models we are approaching to better understanding of how things really are.
The notion of spatial resolution in its common understanding (grid 50×50 m, 10×10 km or 0,5º×0,5º) is not applicable in the concept of the Hydrograph model. The spatial resolution of the model depends on the scale of the watershed under study. Traditionally we use hexagonal grid in our model, though any other configuration may be used. Simulations are conducted for the set of so called Representative points which are located in the centers of the grid cells. Representative point (RP) presents vertical profile of an elementary slope of the watershed with its specific properties. RP is reperesentative for its cell. It is important to note that the characteristics and parameters of RP are not averaged properties of the cell. When transitioning from small-scale modelling (for example, simulation of variable states at soil profile or experimental watershed) to more large-scale applications the principle of generalization is used. Generalization does not mean averaging the conditions but a choice of dominant ones. In such a way, soil profile presents maximum spatial resolution of the Hydrograph model. Minimum spatial resolution depends on the model application and specific tasks.
Maximum time step which may be used in the model is 24-hours. There are no limitations on smaller steps (hours, minutes) if necessary meteorological information is available. However, it is necessary to remember that the Hydrograph model is a hydrological but not laboratory model and adequate time step should be chosen for modelling of processes (for example, it is possible to simulate soil moisture with minute time step but it is not efficient).
The essential fundamentals of physics suggest that the process of runoff formation must be the same in any point of space and the processes are governed by climate conditions, landscape and other watershed properties. So, in the conditions of strong frost and absence of snow cover any soil would start to freeze. However, how fast and deep it would freeze, it would depend on a soil type (its physical properties), its current state (wet or dry) and other factors. So, ideally the parameters of the model should reflect physical properties of landscapes which determine the processes and the model algorithm should be general (single) in processes description. We are seeking for that while developing the model. The necessary condition is the model numerous applications in different conditions of the Earth.
The concept of runoff elements, which is the basis of our approach and conceptual description of water movement within the watershed storages, allows carrying out simulations for basins of any size since the value of basin area is introduced into the calculation scheme (link, read more…). When conducting modelling at small and large basins at the same landscape and climate zone main set of the model parameters stay unchanged and may be ported form one scale to another (see also question 7).
The discussion on advantages of fixed and flexible model structures is in high gear. We think that any model under development is a flexible model as its algorithms are open to refinement and improvement. We think that idea of flexibility in hydrological modelling is being understood improperly: we would not call a model flexible if its structure is selected for each specific watershed and its parameters are intensively calibrated, as such model would be not be applicable beyond the calibration data. At the same time a fixed structure of the model, with developed and confirmed algorithm at many hydrological objects, may become a base for further enhancement which does not have any limitations. Good example of proper flexibility is the Canadian model CRHM. Again, we think that combination of fixed model structure with flexible mind of a researcher is mostly productive.
The Hydrograph Model describes the path of water through a watershed starting from precipitation and finishing with runoff at the outlet. The algorithms of the model include description of all components of the land hydrological cycle: precipitation and its interception; snow accumulation and melt; evaporation from snow, soil and vegetation cover; surface flow and infiltration; soil water dynamics and flow; heat dynamics and phase change in soil layers; underground flow formation, slope and channel flow transformation; and flow discharge.
We constantly continue testing and development of the model in different geographical conditions (please, see the section Projects). In ideal, the model should be ready for reliable application in any point of our Globe, and we are constantly seeking for it.
The model has been developed on the assumption of maximum possible simplicity. Necessary forcing data is air temperature and humidity, precipitation. Information on watershed characteristics is required for assignment of the model parameters, i.e. – a description of relief, soil and vegetation cover, some specific conditions. The most laborious stage of modelling procedure is estimation of the model parameters describing the watershed. Then forcing data is loaded into the model program and simulations are conducted. Then the simulation results are analyzed and compared with the observed data. If necessary, the model parameters are refined (how? – see question 13) and the simulations are conducted again.
The general idea is to develop the Hydrograph Model for its application in any geographical condition and scales – from point to globe. However, as our Research Group is really small some modules of the model do exist but never have been tested.
Those include watersheds in arid climates, areas where groundwater is very close to the surface, glacially dominated or urban-dominated watersheds. Another model limitation is its routing scheme, which is not applicable to rivers subject to backwater effects.
However, the model can produce the runoff to be input as lateral inflow to fully dynamic models where this is a necessity. We hope that the realization of those parts of the model is the matter of nearest future.
Currently as a free-distributed version of the model is not ready, you have to contact us. However, the preparation of the model program which would be available for everybody’s use and detailed instruction on its practical application are in our nearest plans.
We do not accept calibration in the form of automated procedure of parameters estimation (C1 calibration type) and assume its common application to be one of the main barriers in development of modern hydrological modelling. Manual calibration may have different sense. It is not a secret for anybody that each model has a small set of most sensitive parameters which, if are tuned, satisfactory modelling results of runoff at basin outlet may be easily achieved (C2 type). However, there also exist other type of calibration – adjustment of model parameters on the base of observational data of variable states of soil, snow and other components (C3). This is the most useful but difficult and time-consuming type of calibration which requires deep analysis of the data. It allows to systematize the model parameters (=physical properties) by different types of landscapes and with high confidence use them in similar conditions. We never use C1, we are ashamed if we have to use C2 and we get real satisfaction from our work if we have opportunity to use C3.
The Hydrograph model may be used in different practical tasks, for example, runoff forecast, assessment of hydrological alterations due to climate and landscapes changes. Joint application of our model with stochastic model of weather allows get probabilistic estimates of runoff characteristics regardless of the length of runoff observational series.
Currently we are also working with American researchers on looking at the ability of the Hydrograph model in representing soil moisture profiles in areas subject to intensive irrigation, such as California. The results of the study may be used to develop procedures for initializing weather and climate forecasting models, and to forecast soil moisture and temperature weeks in advance in order to help farmers decide on the best irrigation strategy.
In 2010 two monographs on hydrological modelling were published where rather detailed description of the Hydrograph model and its algorithm can be found. However, those publications are available only in Russian. In 2011 the paper on main principles of the Hydrograph model development was published in HP.
Our Research Group is very small, it functions in despite of any logics based on pure enthusiasm of its members. As we do not have any support from our hosting institutes in these activities, our most efforts are spent seeking for grants, projects and conducting them. Therefore preparation of the papers is all the time postponed. However, we clearly understand urgent necessity to publish detailed information on algorithms and results of the model applications. Currently we are working on several manuscripts describing: 1) snow cover algorithm and results of simulations; 2) heat and water balance in different types of soils; 3) results of process simulations at the Dry Creek watershed, Idaho, USA; 4) experience of process modelling in the Yukon River basin; 5) results of process simulation in permafrost environment.
At this moment the Hydrograph model is used only by our Research Group. This is because the model program is still not completely debugged and the instructions on its application which should accompany the program are not completed. The reason for this is that we have very little resources; all work on developing the model, programming, parameters estimation, data preparation, writing reports are conducted by the same people. We have catastrophic lack of time. However we clearly understand the necessity to launch the model into free access and make efforts for approaching this moment.
- The main our dream is that our studies, our work, our efforts would one day become demanded in our own country – Russia, and that the results of our research would be used to serve people in solving different practical tasks.
- In general we think that it is appropriate to develop modelling in three directions – scientific (further development of the model), practical (application of the model in operational practice, necessarily with feedbacks) and educational (training of students and specialists). We dream about developing an interactive training course based on the model.
- We dream that one day the Hydrograph model would be free-accessed. It should be accompanied by the database of the model parameters systematized by different landscapes and detailed instruction on how to use the model.
- Our model should become the base for forecast of catastrophic hydrologic events in the mountains, such as debris flow, breakthrough floods and so on. We dream about that time when we will be able to add those blocks of the algorithm to the model program and test them on real objects. Prof. Yuri Vinogradov is the best specialist on these problems from the times of the former Soviet Union.
- We would like to start working on such problems as possible change of hydrologic regime after forest fire, and many others…
- We dream about many things… and we work hard to bring those dreams into reality…
Our Research Group is open to any collaboration that move forward the development of the model, hydrologic modelling in general or would be useful for solving practical tasks. So we would be interested to apply our model at any new object in scientific or practical purposes, conduct work on comparison of different models or apply our model in combination with other models.