The forecast for the snowstorm that is finally ending across the Washington region was about as complicated as it gets. The scenario involved one storm, with origins in the Pacific Ocean, that would race across the country, die out over Ohio, and then hand off its energy to a secondary storm developing off the Mid-Atlantic coast.

Snow would come in two phases with an icy interlude in between.

So many questions emerged: How much snow would fall in the first phase? When would a “dry slot” stop the snow? Where would the new coastal storm form? When would the second phase begin, and how long would it last? Where would the heaviest snow occur? Which computer model simulations were too snowy or not snowy enough?

Even with all of these unknowns, our forecasts for the event correctly predicted the general evolution of the storm and the amount of snow that ultimately fell. However, the projected snowfall for the first half of the storm was slightly overdone, while the second half was slightly mistimed. It was a case of reaching our destination, but only after veering off course first.

On Sunday, we predicted three to five inches of snow from the storm’s first phase. Mostly two to three inches fell, with a handful of reports of a little more or less. The day started off promising for our forecast, as snow began steadily in the predawn hours, and one to three inches accumulated by midmorning. But then it eased through the midafternoon and picked up only in spurts into the evening.

The snowfall underperformed in the first phase as the parent storm over Ohio quickly faded, allowing dry air to infiltrate the Mid-Atlantic. Our forecast of three to five inches Sunday actually leaned toward the low end of model projections (many showed five to six inches), as we recognized the possibility. In past experience, storms that die out to our west don’t typically produce much more than three inches, as milder air to the south overruns cold air over our area.

But we allowed for that outcome, given model simulations. It turns out those model solutions were more overdone than we thought. That was our first mistake.

As for storm timing, we accurately predicted the pause in the snow that occurred Sunday night before the storm’s second phase, which we knew would be the most challenging to get right.

When storms in the Ohio Valley dissipate and transfer their energy to a new coastal storm, snowfall in Washington is heavily dependent on where the new storm forms and how quickly it develops. In the days leading up to Sunday, model simulations were all over the map as to where the new storm would form and intensify. For our area, the farther south and faster scenarios would carry the best prospects for significant snowfall.

Model simulations Friday were generally promising for our region’s snow potential from the coastal storm, with many projections showing it forming quickly near Cape Hatteras, off North Carolina — an ideal location for producing moderate to even heavy snowfall in the D.C. area.

But Friday night into Saturday the simulations drifted north in their placement of the new low pressure area, first closer to the Virginia Capes, then to just off the Delmarva coast. This had important implications because it put eastern Pennsylvania, New Jersey and southern New York in the best spot to receive the heaviest snow bands.

In drawing our snowfall map, we took into account the possibility that the coastal storm would end up forming too far north to deliver big snows in the immediate area. We know that storms that die off in Ohio usually hand off their energy to coastal storms a little too far north for us to hit the snowfall jackpot. In other words, climatology argued against going for a major snowstorm here.

Typically, secondary coastal storms that produce big snows around Washington derive their energy from lows that dissipate in Tennessee or Kentucky and spin up coastal lows to our southeast.

On Saturday, we predicted just one to three inches for the storm’s second phase, with low to medium confidence, while allowing for the possibility of a significant boom or bust.

When it became even more apparent by Sunday afternoon that the coastal storm would form too far north to produce heavy snow in Washington, we lowered the forecast for Monday and Tuesday to a coating of two inches.

This wasn’t exactly wrong, as most locations in the immediate area got one to two inches from the storm’s second phase, but it turns out to have been an unnecessary adjustment.

While our snowfall forecast for the second phase of the storm turned out to be more or less correct, it evolved a little differently than we expected. We were thinking the most likely period for accumulating snow would be Monday afternoon and Monday night, whereas it occurred much later, during the first half of Tuesday.

When a big, slow-moving coastal storm blows up to our northeast, our snowfall is dependent on narrow bands that wrap around it. Attempting to nail down the timing and intensity of these bands is nearly impossible given the capabilities of computer models. Instead of a forecast, it becomes more of a “nowcast.”

Overall, when you compare how much snow fell over the course of the entire event with what we predicted, it turns out to be a decent match. We exceeded actual snow amounts in locations east and southeast of Washington slightly, where only two to three inches fell over three days. We also wrongly anticipated that areas to the northeast of the city, including parts of Baltimore County, would get more snow than they did.

Parts of northern Maryland near Hagerstown and north of Frederick got a lot more than we anticipated, as they were bombarded by very small scale and persistent heavy snow bands, which were impossible to predict with much skill ahead of time.

Officially, 3.2 inches fell at Reagan National Airport, 5.7 inches at Washington Dulles, and 4.0 inches at BWI Marshall. The highest total in the region was 20.5 inches in Sabillasville, Md., in northern Frederick County.

Several of our television forecast colleagues also made reasonably good forecasts (e.g., see NBC4 and FOX5).

We correctly characterized the impact of the storm, placing emphasis on its long duration and the cold temperatures ahead of the storm, which meant snow and ice in its early stage readily accumulated on untreated roadways. We labeled it a Category 3 “significant” storm but not “major” or “crippling,” knowing it would be spread out over a long enough period for people to navigate through it, with caution.

As the biggest snowstorm in the region in two years, it brought delight to snow lovers and proved not to be a huge inconvenience. Overall, it was a storm that had something for everyone.