North Carolina Drought: Will Our Cities Run Out of Water?

By David H. Moreau
Director, Water Resources Research Institute

Every day during the drought we hear various estimates about how many inches of rain it will take to bring us back to “normal.” The other index is how many days of supply we have on hand. Both of those estimates are useful indicators, but they tell very little about the risk of a city running out of water. As we find ourselves in the middle of a severe drought, no one knows how bad effects of this drought will be on our public water supplies. We can and should turn to records of previous droughts to gain insights, but we are left with the uncertainty of not knowing how bad things will get.

Historical records can provide important clues about the magnitude of what we face. Cities may withdraw water from streams and reservoirs by up to 60 percent of average annual flow, and to sustain that rate, they need to capture a smaller percent of average flow in reservoirs. For example, the Orange Water and Sewer Authority’s source on Cane Creek is expected to yield 55-60 percent of average year-round flow during a repeat of the worst drought of record, but the reservoir needed to sustain that rate captures only 15 percent of the flow from the watershed that feeds it. Sustainable withdrawal for water supply from Falls Lake, Raleigh’s primary source, is about 30 percent of average inflow that is assigned to that purpose. Water supply storage in the lake holds about 19 percent of average annual inflow delivered to it. It does not take normal rainfall to replenish these reservoirs.

What does it mean then that there are 75 to 100 days of supply remaining? That indicator by itself doesn’t tell you much. The Cane Creek example is again a useful example. When it is full, it doesn’t hold but 100 days of its safe yield. On the other hand, the water supply pool in Falls Lake holds about 225 days of dependable withdrawals. If Falls stands at 100 days of supply, it is just below half full. If Cane Creek has 100 days of supply, it is full. These numbers have very different implications for the two supplies.

On any given day, the risk of running out of water depends on how much water is in storage, the rate at which water is being withdrawn, and the rate at which streams are replenishing storage. All of that is highly variable from one water supply to the next, and it changes rapidly with time during the late fall and early winter.

Let’s say, that is the midst of this severe drought, a particular city had 90 days of supply on hand November 1. One way to estimate the risk of running out is to calculate the storage levels on a day-to-day basis for the next several months using current rates of withdrawal and historic sequences of streamflow during the worst droughts of record. In this analysis, we looked forward from Nov. 1 through the end of February. Computations were made at locations of seven stream gages located across North Carolina. Sites were selected using several criteria, including a lengthy record of flows that are subject to very little upstream regulation, watersheds in the approximate range of 100 to 300 square miles, and one or more in each geographic province. The shortest record was 68 years, the longest 82 years. Years in which the November-February flows were among the lowest on record varied somewhat from gage to gage, but the most frequent were 1954, 1955, 2000, 2001, 1941, 1930 and 1980. The lowest six years at each gage were used to calculate if and when reservoirs at those locations would have run dry for a range of initial storage levels and a range of withdrawal rates. Withdrawal rates ranged from 10 percent to 60 percent of mean annual flow, and initial storage levels ranged from 60 to 90 days of supply.

Results of these calculations point out the substantial differences among the seven locations. At one on the eastern slope of the mountains, supplies would not have been depleted under even the worst drought conditions; at a location on the western slope, reservoirs would have run dry for all withdrawal rates under a repeat of either the worst or next worse drought. One Central Piedmont site would never have run out under the worse drought and the highest rate of withdrawal. Sites in the Research Triangle and east showed the greatest vulnerability. A repeat of the 1933, 1941, or 2001 drought would pose very serious problems for water supplies if they are operating with less than 90 days of supply on hand as of Nov. 1 and their withdrawals are more than 20 percent of average annual flow.

We were hoping to find some comfort from this analysis. We didn’t. Repeats of the worst of our previous droughts could have substantial consequences if we don’t get relief over the next couple of months. We don’t need to get to “normal”, but we do need relief. Our best option at this point is to reduce our withdrawals through conservation, reuse, and alternative sources if they can be found.