Stay Informed

Sign up to receive regular updates from the More Profit Per Drop team

Upcoming events

Deep Drainage Mythbusters

Emma Brotherton, Graham Harris, Peter Smith and David Wigginton
For many years it was believed that deep drainage from furrow irrigation was not an issue on the cracking clays predominantly used to grow cotton in Australia. Results from current Cotton Catchment Communities CRC deep drainage research projects and evaluations of commercial irrigations are challenging this belief. At the same we often hear statements providing “evidence” that deep drainage is not occurring when irrigating.

The Cotton CRC Water Team promotes best irrigation practice by providing information, demonstrations, workshops and training. We sought the response of some of the industry´s leading deep drainage researchers on these Deep Drainage “myths” – Anthony Ringrose-Voase and Richard Stirzaker (CSIRO), Willem Vervoort (Sydney University), Des McGarry, Mark Silburn and Jenny Foley (QNRW).

Myth 1: When I irrigate throughout the season the soil seals, reducing infiltration – so how can I have deep drainage?

Once water enters the soil it will fill the soil profile in a non-uniform way, with some rapid preferential flow down larger cracks and pores. Vertosols, commonly used for irrigated cotton production, are often considered to be non-draining but in reality water will drain beyond the root zone, becoming potential ;deep drainage (DD), both during the early stages of irrigation due to preferential flow, and in the later stages of irrigation if too much water is applied. Even the heaviest clay soils still have water moving through them, and infiltration is never zero during irrigation. It takes about 1 hour for water to travel 1 metre down the profile of an initially moist Black Vertosol without large cracks. Drainage beyond the root zone may occur within a few hours of the start of irrigation.

In heavy clay soils the infiltration rate decreases as the soil wets up due to swelling of the clay particles. It is important to separate infiltration (the water moving into the soil from the top) from DD (the movement of water beyond the reach of crop roots). The two may be related; DD often occurs much later and continues much longer than infiltration as water slowly moves through the profile. DD is generally much less than infiltration as part (hopefully most) of the infiltrating water is stored for use by plants.

Some irrigated soils under cotton do have a tendency to seal (close over or form a surface crust) but this is not true for the majority of cotton growing soils. For example, of the 10 sites in a current Deep Drainage project (Project 1.02.04 ´Deep drainage under irrigated cotton´ – Des McGarry), only 3 have a tendency to seal after rain or irrigation. Several soil properties act together to cause this sealing action. Soils with more than 45% clay and high sodium levels (as a proportion of the exchangeable cations) tend to seal as the sodium breaks up soil aggregates, when wetted by rain or irrigation. If fine sand is also present (more than about 20% of the soil mineral fraction) then the tendency to seal is even greater. Low organic matter levels add to the risk. If you have soil with these properties, adding gypsum and/or organic matter (crop residues) may reduce the problem.

The decrease in DD as the season progresses was measured at almost all sites in all years of the project. However, this is not necessarily because water infiltration lessens or because the soil begins to seal over. Rather, as the cotton season progresses, the above-ground biomass and the plant root system of the cotton crop develop and grow larger, and the day time temperatures greatly increase. The cumulative effect is greatly increased evapotranspiration (water demand). Though the irrigation frequency increases to match the increased evapotranspiration (if water is available), the crop demand is so large and daytime temperatures so high that almost all the infiltrated water is rapidly used. SIRMOD analysis, measured across 4 of the current DD trials clearly shows decreased infiltrated depths of irrigation water during the last few irrigation events of the season (see 4th & 5th irrigations in Figure 1). Data from the current DD trials show that, commonly from the 3rd or 4th in-crop irrigation, there is no measured DD.

Figure 1: Infiltration uniformity as simulated from SIRMOD analysis down the furrow on the Goondiwindi site for 6 irrigation events in the 2004-5 cotton season. The modelled amounts of water infiltrated (in millimetres) at each of head, mid and tail locations are shown (in brackets) in the legend.

Some soils used for irrigation do exhibit this “sealing up during the season” behaviour, causing infiltration rate per irrigation to decline through the season and difficulties with subbing-up. A common management response is to run furrow irrigation for longer (sometimes much longer), sometimes generating a large volume of tailwater. However, you still won´t know for sure that DD is not occurring, especially in the first irrigation or in irrigations with long run times, unless you do in-field measurements. It is likely that deep drainage does occur but it may be less than for soils with higher infiltration rates.

Myth 2: Groundwater tables have been falling over the last few years, how can we have deep drainage?

Soils drain and recharge aquifers. Aquifers in turn can discharge to rivers. Most irrigation areas have at least two aquifer systems. The top aquifer is a water table aquifer, and the quality of this aquifer is often marginal to saline. This can be easily derived from the bore logs in the areas. For example, in the Namoi and Gwydir valleys this upper aquifer is called the Narrabri formation. In contrast, most groundwater for irrigation is extracted from the deeper aquifer, which are often much better in quality. In the Namoi and Gwydir valleys this is the Gunnedah formation. The groundwater levels in the deeper aquifers have been falling due to extraction. We don´t know exactly what is happening in the upper aquifer (Narrabri formation) as very few measurements are taken in this aquifer but recent research (Project 1.02.07 ´Hydrological and geophysical characterisation of Palaeochannels in northern NSW´ – Chris Vanags) indicates that the upper aquifer (Narrabri formation) responds rapidly to rainfall and irrigation.

Figure 2: Average daily water levels for the piezometers in the Narrabri formation for 2005 (irrigated) and 2006 (not irrigated) in a field containing a palaeochannel. Plotted data are calculated as the average from the 15 minute observed data over 24 hours. Black bars are recorded rainfall in cm, while the grey bars are irrigation events of approximately 1 ML ha-1. Over the course of the 2005 irrigation season, the water table rose 14 times in the wells inside and below the palæochannel on the SE side (piezometers 2, 3, 5 and 6) (Vanags et al. submitted to Australian Journal of Soil Research).

In addition, groundwater levels decline when the natural outflows plus pumping volumes exceed the inflows plus recharge. Recharge can come from rainfall, irrigation fields and structures, and from rivers and floods, depending on where you are. Thus there could indeed be deep drainage on your farm and falling groundwater levels if groundwater use/outflow exceeds inflows and recharge. The response of water tables to recharge from deep drainage is complicated as there can be a prolonged delay between drainage occurring and recharge to a water table, and this recharge may not be from directly above. If the aquifer was a “closed bucket” under a farm which was entirely irrigated from that aquifer, and 5 ML/ha was pumped for irrigation and 10% went to deep drainage (0.5 ML/ha) each year, then outputs exceed inputs by 4.5 ML/ha and the water level in the aquifer would fall over time. If the aquifer held only 10% water (the rest being solids and water that can´t be pumped), the water table would be lowered by 4.5m each year even though deep drainage is occurring!

Myth 3: My soil moisture measurement tool does not show a change in soil water levels at depth throughout the season – this means deep drainage is not occurring.

This comment is based on the assumption that little change in soil moisture measurements over a period of time shows that water is not moving past the measured depth. Water movement is driven by energy differences rather than by water content. The energy differences are determined by two components, one is the soil moisture (wetter soil is higher in energy than drier soil), the other is gravity (water will more easily move down than move up). So even if the soil moisture difference is negligible (and thus the energy difference is negligible) the gravity can still drive water movement. And this would have very little effect on the soil moisture content. This means deep drainage may be occurring even if the soil moisture measurements do not change over time. The lost water is replaced with water from above. At depth the subsoil is near saturation so it is not possible to detect any increase in soil moisture with deep drainage.

Soil moisture monitoring devices don´t measure drainage. Depending on the type of device, they may indicate if drainage is likely to be occurring. There are two types of devices, those that measure the pressure (energy) at which the water is held within the soil (tensiometers, gypsum blocks, Watermarks) and those that measure the water content of the soil (capacitance devices, TDR probes, soil coring, neutron meters). The first type of device is very useful for indicating whether the soil is “wet” and therefore draining i.e. when the soil water potential is between 0-0.1 bar (0-10 kPa). They give a very good idea of when and for how long drainage occurs, but not how much has drained.

The second type of device measures the amount of water held in the soil, however this does not, in itself, tell the irrigator if the soil is “wet” or “dry”. Generally, a dry soil below drained upper limit (DUL) is not draining, whereas a wet soil above DUL is draining. Soil water content measurements alone will not tell the irrigator this, and related soil information is needed to interpret the measurements. Note: if a device is measuring constant soil water content over time and the soil is above DUL, then that soil is draining, and at a constant rate.

Water content in heavy clay soils may vary only a few percent (2-4%) between saturation and DUL i.e. from maximum drainage rates to no drainage. Such a small change may be overlooked, or lost within the error component of the device. Even when such a change is detected, its meaning is difficult to interpret because soil moisture content does not indicate the absolute moisture status of the soil. Soils can be saturated at moisture contents of 0.4 to 0.6 v/v (the water volume is 40% to 60% of the soil volume), depending on their bulk density. Thus a soil moisture content of 0.35-0.40 v/v would be approaching saturation (and drainage would be occurring) in a dense soil (e.g. a Grey Vertosol) but would be near wilting point in a swelling, high clay, Black Vertosol.

Other reasons why soil moisture monitoring tools (like capacitance and neutron probes) may give poor insight into DD include:

  • The output from these devices shows change in soil water content (between times). That the line does not change, at say 100 cm depth, throughout a season, shows only that there has been no change in soil water content at that depth. However, the water at that depth may well be flowing through the soil, by-passing the root zone as DD. This bypass flow can occur down cracks or old root holes and will not be picked up by a soil moisture monitoring tool.
  • Also, data from Des McGarry´s DD project shows large variation in DD within any one paddock as well as between seasons, and certainly during the season. The largest values of DD measured have been early in the cotton season at the pre- and first-irrigations. Most soil moisture monitoring tools are only installed when the crop emerges – to ensure proximity of healthy plants to the probe. In this way, the early season events (with large water application at the pre-irrigation, low temperatures and small transpirational demand) that contribute greatly to DD are missed.
  • Depending where the soil moisture monitoring tool is located, it may or may not show change (at depth) in soil water content. Unless moisture probes are installed at head ditch, middle of the paddock and tail ditch locations, and monitored throughout the season, including before and after pre-irrigations, it is not reasonable to conclude that there is no DD. Unlike soil moisture monitoring tools, the device used in the Des McGarry´s DD project to provide DD data is a drainage lysimeter. This is not a soil water content measuring device – rather it continuously collects the actual DD water arriving at 150 cm depth. The amounts collected are recorded electronically and the actual DD water is collected routinely (for salinity assessment) during site visits.

Myth 4: My water storage does not leak so deep drainage is not an issue when I irrigate.

This statement assumes that if a storage does not leak then soils irrigated nearby will not lose water through deep drainage. Firstly, all storages leak to some degree, and the greater the head in the storage, the greater the leakage. Secondly, a properly constructed storage with the right level of compaction and construction material is quite different from a field being irrigated – it is not sensible to compare drainage losses from a storage with those from an irrigated field.

A complication is that the water level in storages falls from pumping, evaporation and drainage. Without accurate storage measurement, it is not possible to ascertain how much is due to evaporation and seepage loses. And if it is leaking at less than a few mm per day, it would be difficult to separate seepage from evaporation. So how can you be confident that your storage is not leaking?

In addition, similar to irrigated fields, storages are known to be highly variable in terms of their “leakiness” and where it may occur. In leaky storages, the whole floor of the storage probably does not leak equally. Sand lenses just under the storage floor may give a leakage “hotspot”.

Myth 5: I irrigate efficiently by maintaining high heads and pulling siphons as they come through so no deep drainage occurs.

This statement poses two questions:

  1. How did you measure your efficiency?
  2. How do you compare with other irrigators?

Without measuring irrigation performance of a field it is not possible to know the efficiency with which it is being irrigated. Rule-of-thumb irrigation management strategies applied without measurement may be efficient – but you don´t know for sure.

Two indicators can be used to measure the efficiency of an irrigation event: Application efficiency (Ea – the percentage of water applied that infiltrates into the soil) and Distribution Uniformity (DU – the uniformity of infiltrated depth down the paddock from head ditch to tail drain). These are illustrated in the data given in Figure 1. At the first irrigation, the Ea was 54% with a DU of 71%. That is 54% of the irrigation water applied actually infiltrated, so was available for the crop to use. By the 5th irrigation event, however, Ea was only 35% showing less water available for crop use (from drier soil at time of irrigation, from large evapotranspiration mid-season, as well as less water applied) but DU was 99%, showing an almost uniform distribution down the paddock (though with far less water than at the start of the season). In the same figure, infiltrated depths at all head, mid and tail locations for the pre-irrigation are well over the required depth contributing to deep-drainage. In conclusion, a general result from the current project is that around 70% of measured DD can be attributed to the pre-irrigation and the first two in-crop irrigations.

An examination of 79 furrow irrigation evaluations in Queensland cotton fields by Rod Smith, Steve Raine and John Minkevich found:

  • Irrigation application efficiencies varied widely from 17-100% with an average of 48%,
  • Deep percolation (drainage) losses averaged 42.5 mm per irrigation, representing an annual loss of up to 2.5 ML/ha,
  • Irrigation application efficiencies in the range of 85-95% were achievable by optimising furrow irrigation in all but the most adverse conditions.

Efficient furrow irrigation is possible by fine tuning your irrigation applications to your soils and fields using a surface irrigation evaluation.

One of the key principles of efficient irrigation is that extra water is applied to leach salts which are inevitably added with irrigation water and fertilisers. If no DD occurs, salinisation of the soil would occur. However, DD during rainfall events on irrigated fields will generally provide enough leaching, unless the irrigation water is particularly saline. While some DD will always occur under irrigated fields the key is to make sure it does not occur due to irrigation to avoid the cumulative effect of degrading the environment at both the local and catchment scales.


Deep drainage is a reality in furrow irrigated fields within our cotton-grain farming systems. Some deep drainage is necessary to remove salts from the root zone of our irrigated crops. The extent of deep drainage is dependent upon the soil type and uniformity and the condition of the field being irrigated, the irrigation management practices being used and the seasonal conditions experienced. It is not something that necessarily happens with every irrigation event and the magnitude of it can be managed through improved irrigation practices – many of which are now being implemented within the industry.


This article is a synthesis of responses to the five deep drainage myths by Dr Jenny Foley, Dr Des McGarry and Dr Mark Silburn, NRW; Dr Anthony Ringrose-Voase and Dr Richard Stirzaker, CSIRO; Dr Willem Vervoort, Sydney University; and Thusitha Gunawardena, NRW Queensland. We thank them for their valuable input based on the many years of deep drainage research they have been undertaking through funding from the CRDC, Cotton Catchment Communities CRC and the CRC for Irrigation Futures.

« Back