Workshop 5– Water Accounting

Fact Sheet

LAKE TED HILLYER

Reservoir Information

Location: On the Purple River above the confluence with the Orange River.

Purpose: Water Supply, Flood Control

Lake Data: Based on current sedimentation survey

Feature	Elevation (feet)	Area (acres)	Capacity (acre-feet)
Top of Dam	1095.0	53,300	3,070,000
Top of Flood Control Pool	1085.0	47,182	2,554,000
Top of Conservation Pool	1072.0	39,078	1,994,000
Top of Buffer Pool	1054.0	30,587	1,370,000
Bottom of Conservation Pool	1035.0	22,442	867,000

Operating Zone	Capacity (acre-feet)
Flood Control	560,000
Conservation (total)	1,127,000
Buffer (included in Conservation Zone)	503,000
Inactive	867,000

Model Information

Networks

Basic Network – Contains a single reservoir (Upper) with a diverted outlet

Operations Sets

Basic – Minimum release for water supply

Rules

Min Diversion – Constant year-round release

Alternatives

Basic – Yield Analysis run using Basic network

Simulations

Full Period – Runs the extended historical period (1943 – 2015)

Workshop #5 – Water Accounting

In this workshop you will work with water accounts to enhance the multipurpose reservoir you modeled in the last workshop. You will assign water to a water supply account and use that account to track water supply releases and storage. You will see how this affects the calculation of firm yield.

In HEC-ResSim open the watershed saved as '\Workshop5\Workshop5.wksp'
Open the 'Basic Network' network. The model has the same diversion from reservoir storage that was present at the beginning of the previous workshop. The reservoir has the 'Min Diversion' and 'Hydropower' rules present, which are the same as in the previous workshop.
Now you will create a water account in the model and determine the yield for just that account.
Open the Water Account Set Editor from the Network module.

- Create a new Water Account Set named 'Basic'. This account set will contain the water account you will create on 'Upper Res'.

- On the Reservoirs tab, set 'Upper Res' as an Active reservoir for this water account set. Set the Operation Set for the account set as 'Basic', and make sure the Lower Limit of Con Pool is set as the 'Inactive' zone.

- Switch to the Water Accounts tab. Here you can create a new water account that will live in the water account set you just created, using the Accounts → New menu option. Create a water account to represent the portion of the Con Pool allocated to the water supply purpose, and name it 'Water Supply'.

- The account needs to be linked to a rule on the reservoir that will release water from the account’s storage. HEC-ResSim will track the inflows and outflows to the account and keep a running tally of the account storage. The selected rule will determine the outflow from the account. Select the ‘Min Diversion’ rule. We will not create an account for the hydropower purpose, the remainder of the conservation storage will be not allocated to an account.
- You can specify the size of the account as a percent of the conservation storage or as an absolute volume. For this exercise set the account to 50% of the conservation storage. Click OK or Apply. Apply will save your changes but keep the editor window open. OK will close the window.

Now you need to tell HEC-ResSim to use your water account when it runs. Open the Alternative Editor and go to the Operations tab. At the bottom of the window, choose the water account set that you just created.

Move to the Yield Analysis tab to setup the analysis to find the firm yield of the water account instead of the entire reservoir.
- Change the Yield Analysis Type to Water Account Yield. This will change the available options below.

- Select the 'Water Supply' Water Account and set the Tolerance to 10 ac-ft.

- As before, select the 'Min Diversion' rule.
- This time, set the Convergence Method to Heuristic and Bisection Search. We will look later at the impacts of the choice of method.

Since you created the new water account, you need to give it a starting storage. Have the account start the simulation full. Before you can set a lookback storage for the account, you need to save the changes to the alternative so that HEC-ResSim will recognize that a Water Account Set is selected for this alternative and add a row to the Lookback tab for you. Click Alternative → Save or type Ctrl-S. Now go to the Lookback tab and change the Type for the 'Basic – Water Supply' row to Constant. The Con Pool for 'Upper Res' has 1,127,000 ac-ft of total storage. Set the lookback value to one-half of that or 563,500 ac-ft. If the row is not present in the table, close and re-open the Alternative Editor.

Load the 'Full Period' simulation and make sure the 'Basic' alternative is up-to-date, then run the yield analysis. The run may take longer than in previous workshops due to the addition of the water account.
What is the firm yield of the 'Water Supply' account? Is it equal to ½ of the yield of the full Con Pool? Why or why not?

The yield is 908 cfs. The yield of the full pool was 1508 cfs. Note that the critical period has changed from the third to the first drought in the record. The account yield is more than ½ of the full pool yield. Since not all the Con Pool was assigned to an account, HEC-ResSim created a default account for the remainder. The hydropower rule releases are assigned to this default account, but sometimes the account stays full when the inflow is high enough. Normally an account sized at 50% of the Con Pool would receive 50% of the inflow to the reservoir, but when the default account is full, its inflow becomes surplus and is credited to the 'Water Supply' account. By receiving the full inflow when the default account is full, the water supply account is able to provide a yield larger than ½ of the full pool yield. However, because the minimum allowable storage is reached when the account empties (and the Con Pool is ½ empty), the yield is smaller than the full yield because there is less storage available during the critical drought. The yield is between the full yield and ½ the full yield. We will look at surplus inflows in more detail in the next lecture.

The Storage Yield Analysis Output Summary Report is still available to review the iteration results. The report will show progress towards emptying the 'Water Supply' account with no demand failures. For further insight into the water account, you can examine the output time series directly. View the output using the HEC-DSSVue… option in the Tools menu.

- This will open the simulation.dss file for this simulation, which contains all the run results. Time series for the 'Water Supply' account are named using 'BASIC – WATER SUPPLY' in the B Part. You can plot or tabulate any of these time series.
- Look at the water account inflow: //BASIC - WATER SUPPLY/FLOW-IN//1DAY/BASIC-----0/. Why does the inflow frequently have periods of constant values, sometimes for over 6 months at a time?

The constant values are constant at 908 cfs, the same as the water account firm yield. These periods occur when the account storage is full and the account inflows are greater than the water account firm yield. Each timestep, the account storage is reduced by an outflow of 908 cfs, and is immediately refilled by an equal inflow.

- Look at the water account storage (//BASIC - WATER SUPPLY/STOR//1DAY/BASIC-----0/) and total reservoir storage (//UPPER RES-POOL/STOR//1DAY/BASIC-----0/) together. What do you notice? Since you set the water account size at 50%, is the account storage 50% of the total storage?

The account storage is well correlated with the total reservoir storage. Because the water supply rule has a larger demand than the hydropower rule, the water supply rule affects the reservoir drawdown more. The account storage is less than 50% of the total storage. This is due to the inactive zone at the bottom of the pool. The account is 50% of the conservation storage, but the conservation storage is less than the total storage, even when the reservoir is in the conservation pool, because total storage includes the inactive storage and the conservation storage.

- Look at the reservoir outflow (//UPPER RES-POOL/FLOW-OUT//1DAY/BASIC-----0/) and the water account outflow (//BASIC - WATER SUPPLY/FLOW-DEMAND//1DAY/BASIC-----0/). Is the reservoir outflow the same as the account outflow? Why not?

They are not the same. The hydropower rule also calls for releases, and the reservoir is often at its guide curve, and must pass inflows through to remain there. As long as the inflows are above the water account firm yield of 908 plus the hydropower release, the total reservoir outflows will be larger than the outflows from the two rules.

- Look at the water account inflow and the total inflow (//UPPER RES-POOL/FLOW-IN//1DAY/BASIC-----0/) together. When the water account is less than full, does it receive the full reservoir inflow as credit towards refilling? Why not? Note: this is a nuanced and complicated question but make your best answer.

The answer requires identifying the water account and overall reservoir gains and losses. The reservoir gains are the inflow. The account gains are the account inflow. The difference between the two is caused by the hydropower releases and the losses. Losses for the reservoir are the evaporation (//UPPER RES - POOL/FLOW-EVAP//1DAY/BASIC-----0/). Losses for the water account are not saved in a time series but can be calculated as the reservoir inflow minus the account inflow and the default account inflow. Calculating account losses shows that they are equal to the reservoir evaporation. The water supply account is charged for all the evaporation, not just 50%. This is because the default account is not charged for evap and there are no other accounts to share the losses.

With the account runs taking longer to complete, now is a good time to experiment with the yield analysis tolerances. Go to the Yield Analysis tab on the Alternative Editor. Try increasing the 2 different tolerances one at a time to see how the iteration is impacted. The default tolerances are rather tight, so you can try tighter ones, but you don't have to. It can be easier to understand the results if you set the flow in the 'Min Diversion' rule to the same value before each run. Is one of the tolerances more sensitive that the other? How different is your calculated firm yield with looser tolerances?

There is not a single right answer here, specific results will depend on the exact tolerances chosen along with the starting demand. My experiments showed that generally the storage tolerance was a bit more sensitive than the flow tolerance. Increasing the tolerance enough to cut the iterations in half only changed the resulting yield by < 0.5% showing definite diminishing returns for tight tolerances.

Does switching back to the Bisection Search method impact the run time? What about the calculated yield?

The bisection search significantly increases the number of iterations needed for this simple model without changing the yield. The heuristic search is typically a faster choice with no loss of accuracy.