Catchment grid delineation problem

[FaidessMwale]

I am carrying out watershed delineation with arcmap 9.3.1 and archydro 1.3.

I follow the steps properly from DEM reconditioning and things work just quite fine. The problem arises when i reach catchment grid delineation. although the message says catchment grid delineation is successful, what I get are catchments delineated in some places and empty patches in other places. I thought may be i set the threshold for stream definition too high(4% of maximum flow accumulation) so adopted the default 1% but it does not work.

Can somebody help please; I have been on this problem over a week now.

Thanks

Fai

[MarkBoucher]

what I get are catchments delineated in some places and empty patches in other places.

Fai,
The empty patches could mean you didn't do the fill sinks step. Did you?
Mark

[FaidessMwale]

Hie Mark,

I did; every step was followed diligently.

Fai

[FaidessMwale]

Hie

I repeated the process, I get better results (I dont know why) but still no catchments are delineated in certain areas.

I have attached a map. The black coloured parts represent undefined areas.

Any suggestion  please?

Fai

[LornaMurison]

Catchments are going to be delineated to where your stream segments are, it doesn't look like you have any in the black areas, is this the case?  Are these areas close to the coastline?  If they are, it's possible that flow accumulation in that area doesn't get high enough for streams to be defined because the streams begin close to the shoreline.

[MarkBoucher]

I agree with Lorna.

It could be that the black areas drain away from the catchment outlet. Try the flow direction tool to see if the black areas actually drain via the watershed.

[FaidessMwale]

Hie,

Thanks for replies.

yes, the black section in the northern part surrounds the lake, so it is a flat area. But other black areas are not along the shore line. I checked for stream segments. they are there but as you suggest, flow accumulation may not be enough to generate a stream. I also used a flow direction tool; the flow is not towsrds the catchment but away.

The thing is, the main river originate from a lake and flows for some distance into another small lake from which it resurfaces again and flows downstream (see attachment). I want to calculate the time of concentration for catchment around the main river below the big lake. From Hydrology(if I disregard the big lake), the catchment for the small lake should also be part of the river catchment. Now if i just use catchment A as defined by archydro, am I not disregarding a section of the catchment (which is the catchment of the small lake)? By the way, so archydro does not delineate catchments in low lying areas? what is the way to go around this problem?

Thank you

[LornaMurison]

What you really need to show on that image is the strlnk layer generated by ArcHydro.  ArcHydro will generate a catchment for every segment of that layer.  If you do not have a stlnk segment for the stream flowing into the lake(s) then try decreasing the stream generation threshold.  Have a look at some of the highest flow accumulation values around where the river drains into the lake to get an idea of what value you should be using.
It seems as though the river layer you displayed in the second attachment doesn't agree with the flow direction grid.  It  makes it look as though the big lake at the top is connected to the river in catchment A, but if you look at the rivers shown in the first attachment, catchment A clearly does not drain anything from the lake.

[FaidessMwale]

Thank you Jane.

The lowering of the stream definition threshold improved things tremendously but when I zoom in, in the small lake area (attached), I note that the defined streams follow the natural drainage quite perfectly but there are discontinuities along the main course within the lake area. I repeated this this to the lowest threshold (1 sq km) but with no lack.

[MarkBoucher]

but there are discontinuities along the main course within the lake area.

The discontinuities will appear to be there when zoomed out, but you could possibly see them when zoomed in closer. Its a screen resolution thing.