SWMM-USERS Digest - 11 Mar 2012 to 13 Mar 2012 (#2012-25)

2 views

Skip to first unread message

SWMM-USERS automatic digest system

unread,

Mar 14, 2012, 12:04:10 AM3/14/12

to SWMM-...@listserv.uoguelph.ca

There are 2 messages totaling 144 lines in this issue.

Topics of the day:

1. Evaporation rate and initial event time (2)

**********************************************************
* To sign off, email to: list...@listserv.uoguelph.ca *
* In the body of the message type: signoff swmm-users *
**********************************************************

----------------------------------------------------------------------

Date: Tue, 13 Mar 2012 17:40:13 +0200
From: Gerald Krebs <gerald...@AALTO.FI>
Subject: Evaporation rate and initial event time

I'm in the process of calibrating an urban watershed (~95% impervious) in
SWMM for 6 independent storm events. The watershed is rather small 5.87ha,
but it is subdivided into 690 subcatchments (to allow for estimation of
on-site BMP effects). I'm planning to use multi-rainfall-event
multi-objective optimization for calibration. To minimize the model run time
(the events are from 2 different years) I want to rearrange (create a
continuous series) my calibration events to get them in an as short as
possible time span for one simulation. During this process two questions
were coming up:

1) I used SWMM's routine to compute evaporation based on daily
temperatures. When moving events (4 events stay at their correct date and 2
events move from July 2010 and 2009 to June 2010) the computed evaporation
rates change due to the difference in day length. My idea was to overcome
this issue by computing the evaporation rates first for the original event
dates and use the output as an input evaporation rate (mm/d) in my climate
file. However, my discharge curves differ (between 0.1 and 1 LPS) depending
whether I use SWMM to compute evaporation during simulation or use the
evaporation rates from the climate file. As my model setup is complicated I
used a very simple model (one 100% impervious catchment without any
depression storage and an outfall) to understand (or not ;-)) the issue.
Also this setup shows similar behavior. The water balance in the status
report is the same, losses for the subcatchment are the same, but the
discharge is slightly different (smaller though than in my complex model
setup). Any ideas where that comes from?

2) The second point concerns the initial time periods I need before the
events to maintain the same discharge behavior (initial catchment
condition). I will try to explain the issue using one event. The event dates
6/10/10-6/12/10 (~10mm precipitation). Before the event there was slight
precipitation (~1mm) on 6/10/10. The last rain before that was in the end of
May (~2mm). If I start the simulation on 6/10/10 (including the previous
rain), I get the same discharge as if I start the run 6/1/10. That's still
clear as there was no rain between 5/30/10 and 6/10/10. But when I still
include the rain from end May (simulation start at 5/29/10) I get a
different discharge for the event. The surface depression storage is empty
before the event with any initial run time. Besides surface depression
storage I have only the network and soil moisture to store water. I don't
believe a 2mm event can have an impact on soil moisture for 10 days in June
and affect the runoff (plus my pervious area allowing infiltration is very
small ~5%). I realized that the system storage is never getting empty after
a rain event (in fact it stays at 0.02m3 at the minimum). Also any ideas on
this subject are appreciated. Does somebody have experience with creating a
continuous sequence of independent events in SWMM?

Thanks, I hope I was able to explain somewhat understandable

Gerald

------------------------------

Date: Tue, 13 Mar 2012 10:12:12 -0600
From: Robert Dickinson <Robert.E....@INNOVYZE.COM>
Subject: Re: Evaporation rate and initial event time

Hi Gerald,

Instead of doing this complicated moving of storms have you thought of the possibility using the steady state option to have the original months but eliminate any time period you are not interested in?

Skip Steady State Periods
Checking this option will make the simulation use the most recently computed conveyance system flows during a steady state period instead of computing a new flow routing solution. A time step is considered to be in steady state if the change in external inflow at each node is below 0.5 cfs and the relative difference between total system inflow and outflow is below 5%.

Regarding your two questions:

Your losses are the same but your final storage may be different in the segmented storms than in the continuous storm events. The antecendent moisture conditions are certainly different.

It might help your segmented storms if you have a long interevent time between the storm events.

I am sure that there are many other ways to do what you suggest.

Robert Dickinson
Innovyze Inc.
9340 Pontiac Drive Tel: 813-712-0664
Tampa, Florida USA 33626
robert.d...@innovyze.com
www.innovyze.com

-----Original Message-----
From: SWMM-USERS [mailto:SWMM-...@LISTSERV.UOGUELPH.CA] On Behalf Of Gerald Krebs
Sent: Tuesday, March 13, 2012 11:40 AM
To: SWMM-...@LISTSERV.UOGUELPH.CA
Subject: [SWMM-USERS] Evaporation rate and initial event time

I'm in the process of calibrating an urban watershed (~95% impervious) in SWMM for 6 independent storm events. The watershed is rather small 5.87ha, but it is subdivided into 690 subcatchments (to allow for estimation of on-site BMP effects). I'm planning to use multi-rainfall-event multi-objective optimization for calibration. To minimize the model run time (the events are from 2 different years) I want to rearrange (create a continuous series) my calibration events to get them in an as short as possible time span for one simulation. During this process two questions were coming up:

1) I used SWMM's routine to compute evaporation based on daily
temperatures. When moving events (4 events stay at their correct date and 2 events move from July 2010 and 2009 to June 2010) the computed evaporation rates change due to the difference in day length. My idea was to overcome this issue by computing the evaporation rates first for the original event dates and use the output as an input evaporation rate (mm/d) in my climate file. However, my discharge curves differ (between 0.1 and 1 LPS) depending whether I use SWMM to compute evaporation during simulation or use the evaporation rates from the climate file. As my model setup is complicated I used a very simple model (one 100% impervious catchment without any depression storage and an outfall) to understand (or not ;-)) the issue.
Also this setup shows similar behavior. The water balance in the status report is the same, losses for the subcatchment are the same, but the discharge is slightly different (smaller though than in my complex model setup). Any ideas where that comes from?

2) The second point concerns the initial time periods I need before the
events to maintain the same discharge behavior (initial catchment condition). I will try to explain the issue using one event. The event dates
6/10/10-6/12/10 (~10mm precipitation). Before the event there was slight precipitation (~1mm) on 6/10/10. The last rain before that was in the end of May (~2mm). If I start the simulation on 6/10/10 (including the previous rain), I get the same discharge as if I start the run 6/1/10. That's still clear as there was no rain between 5/30/10 and 6/10/10. But when I still include the rain from end May (simulation start at 5/29/10) I get a different discharge for the event. The surface depression storage is empty before the event with any initial run time. Besides surface depression storage I have only the network and soil moisture to store water. I don't believe a 2mm event can have an impact on soil moisture for 10 days in June and affect the runoff (plus my pervious area allowing infiltration is very small ~5%). I realized that the system storage is never getting empty after a rain event (in fact it stays at 0.02m3 at the minimum). Also any ideas on this subject are appreciated. Does somebody have experience with creating a continuous sequence of independent events in SWMM?

Thanks, I hope I was able to explain somewhat understandable

Gerald

------------------------------

End of SWMM-USERS Digest - 11 Mar 2012 to 13 Mar 2012 (#2012-25)
****************************************************************

Reply all

Reply to author

Forward

0 new messages