Calculating flow accumulation in area?

[TolgaPilevneli]

I want to calculate the total flow in an administrative area which is independent of the basin boundaries. I used the DEM file from HydroSHEDS and so far I have followed these steps:

1. Spatial Analyst -> Hydrology -> Fill
2. Spatial Analyst -> Hydrology -> Flow Direction
3. Spatial Analyst -> Hydrology -> Flow Accumulation
4. Spatial Analyst -> Math -> Logical -> Greater Than
5. Spatial Analyst -> Hydrology -> Flow Length
6. Spatial Analyst -> Hydrology -> Stream Link
7. Spatial Analyst -> Hydrology -> Stream Order (Strahler)
8. Spatial Analyst -> Hydrology -> Stream to Feature

The figure below shows the results of my work. Red lines are watershed boundaries and yellow lines are administrative zone boundaries. There are more than 80 zones. I have highlighted one of them for better visuals.

Now since I know the total amount of water available in a watershed (flow measurements), I want to estimate the relative amount of water accumulating in each of these administrative zones.

Flow accumulation already calculates a value at point O. If I can calculate the relative share of A, B, C and D, I can calculate the amount of water stream A accumulates. However, I still need to calculate the accumulation at points E, F, G, H and I, which are from another watershed, in order to calculate the total amount of water in the yellow zone. A piece of information such as 40 grids flow into point O from point A and 50 grid flow into point O from B+C+D is enough for my needs. And for the other basin the accumulations in E+F+G+H+I relative to that basins flow grids.

Is there any automated way to calculate these values?

[MatthewDriscoll]

It is not exactly clear what you are asking, but you could probably start by calculating the watershed of each point to get a better idea.   Fill -> Direction -> Accumulation -> Snap Pour Point (these will be E, F, G, H, I) -> Watershed.

[MartinPflanz]

Tolga - There are a few parts to the puzzle that I am wondering if they may be key to answering your question - depends on what the main goal is in your task.  Storm events in a watershed follow a volumetric wave, somewhat bell shaped curve - a hyetograph - so at any given time the flow leaving one stream may to a connecting point may have a different peak time than another stream based on transit time (length) in the watershed.  This may complicate some of the manual calculations.  Add to that there, can be different storm events to account for 100 yr, 50 yr etc, which I bet you have looked at already - each has a different hyetograh.  

Putting that together, streams E-I would each have a different peak point in time - and these are streams leaving the yellow area you have marked and differing peaking time steps.  A-D have flow criss-crossing the yellow boundary and again at any given time have a different stream flow peak than each other - largely flowing into the yellow zone.  Calculating the flow accumulation at point O seems straight forward with a hydrologic model, but I support you in that the calculations you are trying to make are difficult.  

If this project is still going/in need, could you reply back with what you are doing with or need to do with the accumulation?  designing or sizing something?   Are you just focusing on point O or all yellow zones in this area?

(very sorry - late reply just came across your original post)

Respectfully,

Martin

[TolgaPilevneli]

Martin thank you for your interest. I know that the flow of each stream could differ greatly depending on the precipitation/surface permeability/snow cover etc. Fortunately, I only need a comparison between each stream at steady-state conditions without considering other parameters.

I am trying to calculate the annual amount of water availability in an administrative region in Turkey. If I can calculate it, I will compare the values with the agricultural water footprint. The problem is just like in Spain, we also keep records of agricultural production (total area harvested, total yield, cost/profit, total irrigation) by administrative boundaries (City/province/region). Also, we know the exact amount of water availability in each watershed. However, the administrative boundaries and watershed boundaries are different than each other. 

So I had two choices. The first one is to convert provincial agricultural data into watershed data or the second one is to convert watershed water availability to administrative water availability, which I chose the latter. Converting agricultural production is a lot harder than calculating water availability in each administrative zone in my opinion. To do such a thing, the exact locations of all agricultural productions and amounts should be known.

Since the water availability for each watershed is already modeled in Turkey (also considering the climate change scenarios until 2100), I don't have to calculate it again. What I need is the relative amount of water running from each stream that flows into the yellow boundary. For example, the watershed with points A, B, C, D is a closed watershed. I know the annual water availability. If I can somehow calculate the proportional flow of A, B, C and D, I can calculate the amount of water supplied to the yellow boundary by stream A.

I am now using a simpler methodology which is a lot easier to calculate but inaccurate. I calculated the total area of each administrative zone within watersheds and distributed the water availability proportional to their areas. So, if you have a more accurate and simple solution, without requiring extra modeling and data, I want to hear about it.

Regards,

Tolga