Rainfall, Runoff and Infiltration Re-visited

- By:

Courtesy of Courtesy of Innovyze


The flow variations in sewers have been modelled for several years with a comparatively high degree of accuracy as far as the diurnal variations and responses to climatic conditions are concerned. However, until comparatively recently the variations in flow due to fluctuations in groundwater levels, soil storage capacity and wetness of the catchment have only been modelled in basic terms. For the past decade it has been possible to use runoff models that respond to the increasing wetness of the pervious surfaces in a catchment during a storm event. However, due to the absence of suitable initial wetness parameters to use with synthetic design storms this runoff model has not been widely used. This runoff model first released in 1991 is still referred to as the ‘New’ UK Runoff Model; perhaps the time has now come to revise this name.

In the days of WASSP and WALLRUS most modellers were familiar with the limitations of the runoff models in use and various techniques were developed to ensure that the model parameters stayed with certain limits within which the runoff routines were sufficiently robust. With the introduction of Hydroworks and Infoworks many of the techniques (or fudges) have fallen out of use with many of the younger generation of modellers falling into the same traps, which their elder colleagues fell into ten years before.

Over the past few years there has been an increasing need to simulate long periods with simulations of a period of several months is not now uncommon. Some of the runoff routines previously used are not suitable for such extended simulation periods; chiefly because of the way in which the interaction of rainfall and the soil storage is simulated.

The purpose of this paper is to review the techniques that are available nowadays for simulating rainfall, runoff and infiltration. Some of the techniques are only suitable for certain limited applications whilst others can be more universally applied. This paper will also explore the suitability of the different techniques for different applications.

Before considering the different techniques available and their suitability it is useful to review the physical processes which occur that we are trying to replicate.

The Physical Processes

The Water Cycle requires no explanation here. It can be considered to start with evaporation from the sea creating clouds of water vapour which then move over land and deposit their water load as rain. In some parts of the UK it rains considerably more than in other parts due to a number of factors including topography, altitude, prevailing winds and the distance from the sea. The rainfall lands on the ground where some of it is immediately lost (the “Initial Losses”) due to evaporation (especially off warm surfaces), wetting of surface layers or dust, absorption into shallow surface layers and the filling of shallow depression storage (from which it later evaporates).

If the rainfall exceeds the initial losses a proportion of the rainfall creates runoff whilst the remainder soaks into the ground and contributes to the soil storage. The proportion that is directed to runoff depends upon a number of factors but the chief one is the nature of the surface. An impermeable surface such as roads or roofs cause a high proportion of the rainfall to contribute to the runoff whilst a recently ploughed field may create no runoff with all of the rain soaking into the soil.

The degree of wetness of the surfaces varies throughout the period of the storm. Impermeable surfaces tend to become fully saturated very quickly and thereafter all of the rainfall is turned into runoff though some of that runoff may go to permeable areas and may not enter any drainage network. Once these impermeable surfaces have been wetted the percentage runoff does not vary greatly.

Permeable surfaces react differently. As the storm progresses the upper layers of the soil become wetter and wetter and when the rainfall exceeds the rate at which it can soak into the ground the rainfall is turned into runoff and when the rainfall intensity drops below the soakage rate the runoff ceases even though rainfall may still continue. Therefore the percentage runoff varies throughout the duration of the storm.

Runoff from impermeable surfaces is routed via guttering, rainwater pipes and gullies into the sewerage system. The extent of impermeable surfaces that drain to sewerage networks varies considerably depending on a number of factors. Areas that have good soakage characteristics frequently have roofs and paved surfaces connected to soakaways. In many areas the roads and roofs drain directly into the sewerage systems. This is not always the case with footpaths, especially where there is a grass verge inbetween the footway and the road – in these cases the impermeable surfaces may drain to permeable areas. Therefore most sewers with any impermeable surfaces connected, exhibit a quick response to rainfall with a peak flow occurring a short time after the peak rainfall.

Rainfall, which soaks into the ground, is directed into the ‘soil storage reservoir’ but when the soil reaches a certain saturation threshold (the percolation threshold) water starts to percolate downwards. A proportion of this percolation flow might infiltrate directly into the sewer network whilst the remainder penetrates deeper into the groundwater storage reservoir. The infiltration flows into the sewers are therefore not dependant upon the groundwater table rising up to above the invert level of the pipelines.

When the groundwater reaches a certain level the water will start to flow through the ground towards rivers, streams and the sea. Where sewers are constructed below the level of the groundwater there can be continuously infiltration into these sewers at any points of defect where the pipelines are not adequately sealed.

Infiltration flows into sewers can therefore either be constant (with slight variations) or they can be in response to rainfall. For simplicity these two different types are termed “Groundwater Infiltration” (GI) and “Rainfall Induced Infiltration” (RII). To complicate matters further, in coastal areas there can also be a groundwater response to tide levels. This variation in groundwater levels causes a variation in the infiltration rates into the sewers and it is best to still consider this as “groundwater infiltration” though it will now have a cyclic variation rather than remain constant.

Flows in sewers therefore stem from a combination of sources:-

  • Domestic foul flows
  • Commercial foul flows
  • Trade Effluent discharges
  • Groundwater Infiltration
  • Impermeable Runoff
  • Permeable Runoff
  • Rainfall Responsive Infiltration


A major step forward with the modelling of rainfall was made when the ‘Wallingford Procedure’(1)was introduced in 1983. Much of the pioneering work was undertaken in the development of the “Flood Studies Report”. The derivation of the parameters needed to generate realistic synthetic rainfall profiles for any point in the UK is a simple exercise using four maps of the UK with the relevant parameters plotted.

Since the introduction of the Wallingford Procedure and the early versions of WASSP and WALLRUS the modelling community has embraced the simplicity of the approach which allowed design storms to be generated very quickly and appropriate catchment wetness indexes also determined very quickly. There were however a number of problems with the methods used for the generation of the rainfall profiles which caused the design storms to not adequately match the local variations which occurred with rainfall.

In 1999 the “Flood Estimation Handbook”(2) was published with a CD-ROM(3) that allowed rainfall depths to be generated far more quickly and with greater accuracy, taking account of local variations. The Author presented a paper(4) at the 2001 Spring Meeting about the use of FEH rainfall in modelling.

Modelling using FEH rainfall is gaining greater acceptance in the industry though in many areas the regulatory framework is tending to inhibit greater use.


Rainfall runoff modelling is extremely complex and has been the subject of much academic research over past decades and indeed is continuing today in many establishments around the world.

The software produced by Wallingford Software contains a variety of different runoff models which can be used. The variety of models available can be daunting and it is not surprising that many modellers do not fully understand all of the models available. The more commonly ones used are reviewed in the section below and are also summarised in Table 1 below:-

Customer comments

No comments were found for Rainfall, Runoff and Infiltration Re-visited. Be the first to comment!