More Numbers

Moore St channel in detail

What is the capacity of the channel now? What is the maximum possible it could be? How obstructed is it? How much would clearing change things? What is a realistic expectation of clearing?

The flow conditions in an open channel are related by Manning's equation:

Q = AR^2/3S^1/2/n

where;

• Q = channel discharge (m³/s)
• A = flow area (m²)
• R = hydraulic radius area/wetted perimeter (m)
• S = slope of hydraulic grade line
• n = Manning's roughness coefficient

Commonsense in symbols: the amount of water that can flow is proportional to the width and depth of the channel, to its flatness or spread, and to the slope it is flowing down, but reduced by any obstructions it meets.

Normally this equation would be used to find flow volume, Q cubic metres per second (m³/s), by plugging in numbers from measurements and tables.

But since we already have pretty good estimates for flow and measurements of cross section area we can use the equation “backwards” to derive a Manning coefficient of friction, n, or how obstructed the channel is, an index number to compare with other known cases - a reality test.

The channel just below Water St is well defined by the newly exposed rock wall (see above), and provides a straight forward point to examine. It also happens to be a point of key interest and differing views.

The channel wall built in 1862 is 2m high and the channel width is 18m giving a capacity flow area, A, of 36m². (this also happens to be the same flow area as the culverts at the Castlemaine Rd bridge, which we will return to below)

The wetted perimeter, R, is the bed of the creek and its banks, 18+2+2 = 22m.

The gradient or slope, S, in this upper section is somewhat steeper than the average 2 percent through the town, and measured at 3.13 percent.

Thanks to observations on the 13th of Jan we know that the channel itself can currently handle at least 30m³/s, and possibly as much as 60m³/s.

From these figures we can derive Manning's roughness coefficient for the channel, and see how it compares to other surfaces. The derived roughness value, n, for this volume range is between 0.073 and 0.145.

A general guide is:

• Concrete pipes n = 0.013
• Concrete channels n = 0.015
• Road Reserves (flow in road and verges) n = 0.025
• Grassed reserves/floodways (regularly mowed) n = 0.035
• Rural floodplains in cleared paddocks n = 0.05

At 100m³/s it fell to n=0.0435.

These values for n look reasonable; we can see the bed of the creek is obstructed by vegetation and scree that has been carried down, and it makes sense that the resistance is less when the creek floods onto the roadways.

But even a figure of 0.04 is quite a high flow resistance, and a number to seriously beat if flooding is to be prevented.

What if...?

We can also apply Manning's Equation “forwards” in the more normal manner to see “what if” the channel were replaced by a concrete channel of the same width and depth.

This should provide a maximum flow figure for an “ultimately” cleared channel, a smooth concrete surface, and is also the absolute maximum volume of water that can pass down the Moore St channel of that width, depth, and fall without flooding.

For n=0.015 (concrete channel 18m x 2m x 3% gradient);

Q = (A*(R^(2/3))*(S^1/2))/n
Q = 294.43

Roughly 300m³/s, absolute maximum.

This is about three times the capacity of the Water St bridge to admit water to this section.

This volume implies an upper velocity of (294/36) or over 8m/s, however best practice calls for a maximum water velocity of 2-2.5m/s to prevent damage to the structures. In practical terms this limitation sets an upper capacity for the channel of around 100m³/s for a 3m/s flow (and the erosion at Hammon Park is an object example of how much damage even 3m/s can do in a few hours).

More realistically, if the growth and scree were removed from the bed and sides of the channel it would have a friction more like 0.025 to 0.035 giving a capacity of between 126m³/s and 177m³/s.

This suggests that the Moore St channel, as built in 1862, would have carried the 100m³/s event on Jan 14th without flooding.

So clearing the channel itself should make a very large difference to its flow capacity.

It also means that the rows of trees beside the channel play no role in creek flow - they are out of the flow until it floods, then the water flows freely down the creekside roadways. They may even serve to protect the Albert St properties from impact by larger floating debris (e.g. firewood rounds).

In a cleared state the channel could easily deliver more than 100m³/s at the Castlemaine Rd bridge. This implies a velocity of 3m/s which is actually faster than desirable in an earth channel, particularly if erosion of the west side levee is to avoided.

* * *

The Bottom Line

Then we come to the Castlemaine Rd bridge itself.

On 13th Jan it demonstrated a flow capacity of only 30 to 40m³/s, and is topping the road at around 60m³/s. Most of the excess from the 100m³/s flow on 14th Jan went via Moore St and North Pde, while the remainder flowed over Albert St.

Culvert design flow rates are specified on the basis that the water can fall freely from the outlet, that it is not flooded. That was the case for the Clunes Rd bridge which had a clear-tailwater, but the Castlemaine Rd bridge tailwater is obstructed by the Clunes Rd bridge headwater.

Because the Clunes Rd bridge is so close downstream if it gets even slightly obstructed the effect is reflected upstream as increased headwater level which also happens to be the tailwater level for the bridge above, reducing its flow capacity. Water then backs up above the Castlemaine Rd bridge because its tailwater is obstructed by the headwater of the following bridge.

And when the waters went down, look what appeared upstream from Clunes Rd bridge...

It is normal practice to design allowing for a 1 in 10 year event to overtop such a bridge by as much as a metre over the top of the culverts, given certain conditions. One of these conditions is local suitability, but with the top of the bridge culvert about level with the floors of several nearby historic houses anything above the top of the culvert is unacceptable.

This bridge should have been designed to cope with a 1 in 10 event fully within the culvert capacity, but this would have required designing for a 1 in 100 overtopping, and providing a considerably larger flow area for both bridges.

Compared to anywhere else along the creekline through town, these two bridges stand out at the major impediment to creek flow by a large margin, a factor of two or three times.

The contribution of sundry Melbourne Water publications on drainage standards and flood control is acknowledged.

Back