Surface Flux Oscillations

A discussion forum for old Hydrus-1D users.
Post Reply
brennan4@umbc.edu
Posts: 8
Joined: Wed Sep 05, 2012 5:25 pm
Location: USA
Location: Baltimore
Contact:

Surface Flux Oscillations

Post by brennan4@umbc.edu » Tue Oct 09, 2012 5:37 pm

<b>Background:</b> I am running a 24 hour “design storm” through a uniform, 100cm soil profile with hydraulic parameters (Qr / Qs / Alpha / n / Ks / I) of (0.0454 / 0.3518 / 0.03 / 1.3427 / 0.127 / 0.5). The storm is an NRCS type II storm, delivering a total of 10.3 cm, with 80% of rainfall occurring between hours 7 and 17. Time units = hours, and precipitation data is in 6 minute increments (240 total), no ET data.

<b>Problem:</b> The Surface Flux plots show instability. The sharp cusp at ~6.5 hours reflects a flux (infiltration) that matches the precipitation input (until ~6 hours), then a gradual leveling to the saturated conductivity of 0.127 cm/hr (by ~12 hours). The instability when time > 10 hours is what is disconcerting. Per Dr. Jirka’s suggestion in a previous thread, I lowered the water content tolerance which reduced the oscillations
Link: http://i.imgur.com/FfoGb.jpg

The oscillations also appear in my observation nodes (placed at 0,1,2,4,6,8,10 cm)
Link: http://i.imgur.com/RgxMJ.jpg

I have adjusted the minimum timestep and time step multiplication factors but have been unable to remove this oscillation. Adding a surface layer of 0.1 - 0.2 cm before runoff drastically reduced the oscillations late in the simulation, but they are still present (figures posted include this surface layer).

Are there any recommendations for reducing/removing this instability? Can anyone provide insight on the nature or interpretation of these instabilities? I'm hesitant to move forward or use my results until I can understand what is happening here.

Thank you!

Post Reply