NATCHEZ — The Mississippi River’s enormous swath of water has consumed conversations for weeks.

Water levels dropped below 60 feet on the Natchez gauge Saturday but are still nearly 12 feet above flood stage.

The gargantuan volumes racing past the Miss-Lou have fascinated hundreds if not thousands of river watchers who have pored over a myriad of websites, dazzling watchers with river levels, forecasts and flow rates.

River levels were the easiest to track, but the flow rate — or the volume of water flowing past a given point — is interesting scientifically and from curious fascination.

But figuring out exactly how much water is flowing past can be tricky, experts say.

Keen river statisticians may have noticed that early estimates of the volume of water expected to pass Natchez at any given point recently disappeared from National Weather Service websites.

No grand conspiracy is in the works, says Jeff Graschel, the service coordination hydrologist at the Lower Mississippi Valley Forecast Center in Slidell, a wing of the NWS that studies and predicts Mississippi River levels.

The early numbers posted on the NWS website were simply inaccurate, he said.

“It was really a source of confusion,” he said. “In light of that, we took the information off.”

Trying to understand exactly how hydrologists create forecasts is a study in why, well, most of us are not hydrologists.

Flow rate forecasts are derived with the use rating curves created using both observed data and historical data. Forecasts take into account the topography of the river’s channel — how deep and how wide it is at various points.

Ratings curves tie together the relationship between a particular river stage and a flow rate of the water, based in part on historical averages.

If nothing ever really changed in the river channel, one could logically assume that when the river reaches X level, the amount of water flowing past will be X, based on what the river had done in the past.

But it’s not that simple, Graschel said.

“For each stage you get to, there’s a flow measurement for it,” he said, using a Vicksburg river level as an example. “But at 56.3 feet today, the flow is different than it was in 2008 or in 1973. It changes from event to event.”

That can be caused by a number of variables, including changes in the river’s channel over time that can affect the volume of water than can fit in the channel.

Or, as in the most recent case, river levels were so high that they were off the historically based charts, making the old rating curve inaccurate.

“When you’re at the high end, you don’t get any measurements at that end at all, you don’t have any historical evidence to go on,” Graschel said.

In addition to modeling, crews from the U.S. Geological Survey use acoustic Doppler radar to determine physical flow rates.

That data is shared with the National Weather Service and the U.S. Army Corps of Engineers.

When the actual data recently found in the boat surveys greatly differed from the forecast models, the NWS forecast was pulled.

“The (forecasted) flows were just not representative of what was going on,” he said. “We felt it was better to take that off so we wouldn’t cause confusion.”

Corps of Engineers public affairs chief Justin Ward said their data showed the peak flow in Natchez occurred May 20, when 2.227 million cubic feet per second flowed past. That’s approximately 16.65 million gallons per second.

As the waters recede, so does the volume racing past. By Tuesday, the level had dropped to 2.135 million cubic feet per second, or just less than 16 million gallons per second.