We use cookies to make your experience of our website better. You can delete or block cookies, but some parts of this website won't work without them. By using this website you accept our use of cookies.

maspiers

29/09/2016 11:10:39

We've found that we get different rainfall profiles depending on whether these are produced in FloodModeller (say, using an ReFH boundary) or in other software (InfoWorks ICM and ReFH2). 

This graph comapres the results for 1:5yr 1hr summer event, scaled to give a total rainfall of 1, for profiles from InfoWorks ICm, FloodModeller using a ReFH boundary unit, ReFH2 using FEH99 parameters and ReFH2 using FEH13 parameters.

http://imgur.com/56fjm5F

Anyone have any idea why these are different?

bpark

29/09/2016 13:19:40

Hi Maspiers

I work for WHS, the developers of the ReFH2 software, and i have looked into this with some of my collegues this morning.

Posted by the WHS software team

There is an typographical error in the equations for calculating storm profiles within the FEH Supplementary Report No. 1 “ The revitalised FSR /FEH rainfall-runoff Method” (Page 30).

The equation for z is reported as z=xb whereas it should have been reported as z=x^b.  The implications of this for calculating the proportional depth of rainfall, y, falling in the temporal proportion, x, of the total duration centred on the peak is summarised in the table below.  As can be seen from the table the profile generated within ReFH2 is correct (for a value of x=1, y=1). If the equation for z, as reported in the supplementary report is used this leads to an error in the calculation of y. This is more marked for summer than winter.  Please note that if you are using ReFH2 you are entitled to support as part of your lease.

  

ReFH2

FEH Supplementary Report 1

Winter

Summer

Winter

Summer

b

1.026

0.815

b

1.026

0.815

a

0.06

0.1

a

0.06

0.1

x

y

y

x

y

y

0.1

0.248

0.330

0.1

0.267

0.190

0.2

0.444

0.514

0.2

0.467

0.348

0.3

0.594

0.642

0.3

0.616

0.478

0.4

0.709

0.738

0.4

0.729

0.587

0.5

0.797

0.811

0.5

0.813

0.676

0.6

0.863

0.868

0.6

0.876

0.751

0.7

0.913

0.912

0.7

0.923

0.812

0.8

0.950

0.948

0.8

0.958

0.864

0.9

0.979

0.977

0.9

0.985

0.906

1

1.000

1.000

1

1.005

0.941

 

 

 

Konrad

Flood Modeller Support

29/09/2016 16:10:53

Hi,

That the results differ slightly is no surprise - I know there is a functional approximation to the standard rainfall profile shapes which was published shortly after FSR, whereas Flood Modeller uses tabulated values from the FSR graph (for large time intervals; for finer intervals it uses the same formula as it's smoother).

What is suprising is that the Flood Modeller results appear consistently higher, so it's difficult to think that the total rainfall in each can be 1 - Just wonder if we're comparing exactly the same quantity, e.g. is any seasonal correction factor or ARF applied to either profile? *Do* the rainfall depths all add up to 1?

I don't think any of the DDF models (1999 or 2013) should affect the profile shape, since you've normalised it to a total depth of 1.

(Not sure what the blue line uses, but looks like it's not centred on a single peak as per the others so presume it's using an even number of data intervals, whereas the others are odd)

Regards

Konrad

Konrad

Flood Modeller Support

29/09/2016 17:04:17

For clarity, the "functional appoximation" I mention is exactly that which bpark refers to, thanks!

Sign in to create or reply to forum posts.

Privacy & Terms

Copyright