Providing us with services from forecasts to life-saving warnings, the National Weather Service is easily the most important and effective agency in the federal government. Though effective, it suffers from some fatal flaws, the agency must be rebuilt to save itself and protect the public.

Counties are awful.

The major problem facing the National Weather Service's ability to more efficiently serve the public stems from a restraint on all government agencies, for better or worse: jurisdiction. In this case, counties.

Americans love the idea of small, decentralized government. The smaller the better, from state governments down to neighbors micromanaging the color of each other's shutters through the power of homeowners associations. On a day to day basis across the country, the county is probably the most commonly used unit to communicate location.

For many of us who don't live in major cities like New York or Los Angeles, we use counties to tell each other where someone lives or where a business is located. Instead of living in Horsepasture, you live in Henry County. Citronelle is up there in northern Mobile County. A good number of William Faulkner's works were set in the fictional Yoknapatawpha County. Instead of using small towns or obscure landmarks few have heard of, counties make it easier for us convey location to one another.

As such, the National Weather Service uses counties to break down almost all of its products. It's quick, easy, and convenient. Products like tornado watches or winter storm warnings are issued on a county-by-county basis. Up until a few decades ago, even forecasts were issued on a county basis. And each National Weather Service office is responsible for a small area of the country outlined by—you guessed it—counties.

Weather Forecast Offices

The National Weather Service devolves its powers much like the country it serves. The agency is broken down into six regions covering the United States, and each of those six regions (central, eastern, southern, western, Alaska, and Pacific) are further broken down into individual Weather Forecast Offices (WFOs).

There are 122 National Weather Service offices in the United States—116 covering the contiguous United States, three in Alaska, and one each in Hawaii, Puerto Rico, and Guam. The area that each WFO serves is known as a County Warning Area (CWA). The size and population that each WFO covers varies greatly around the country—NWS Key West only covers the Florida Keys, while NWS Albuquerque covers an enormous swath of New Mexico.

Each WFO is responsible for issuing a multitude of products for residents within its boundaries. When you go to weather.gov to retrieve your local forecast, that was produced by your local WFO. When you hear a tornado warning come across the radio, that was also issued by meteorologists at your local WFO.

It wasn't always this clear-cut...

The agency underwent a sweeping modernization effort in the early 1990s with the goal of making the organization smaller and more efficient in its operations. Prior to the modernization act signed by President Bush in 1992, there were hundreds of NWS offices around the country. The above map shows the locations of the offices both before (left map) and after the modernization (right map).

In the first map, each red dot shows one of the 204 National Weather Service field offices, and each green square shows one of the 50 forecast offices around the country. The sheer number of offices around the country was inefficient and costly, so the agency close and consolidated those hundreds of offices into just 122.

The agency has come a long way from its scattered past, but the way the agency is currently set up is still inefficient and even dangerous at times.

The Problem(s)

National Weather Service meteorologists are top-notch and are both excellent professionals and excellent people. However, their forecasting abilities (and the public's confidence in the agency's products) are hampered by our near-religious adherence to the concept of counties.

Nature does not respect political boundaries. They are made-up constructs that exist to help us maintain order and control. We draw them with little regard to nature and often with little regard to each other. As such, it creates some major problems when large-scale weather events cross from one artificial political entity to the next.

Using counties is an antiquated way to disseminate weather information in the digital age. It was okay back before the internet era, but now that Uber can track you down to within feet of where you're standing, the National Weather Service should have absolutely no problem breaking down weather information to a sub-county level.

Counties create major problems for communicating the weather. Counties that have a large east-to-west extent can experience enormous differences in conditions from one end to the other. The western side of a county could receive a thick blanket of snow, while the eastern side has just a rain/snow mix. Tornadoes often cross from county to county and from one WFO's area of responsibility to the next, providing a logistical nightmare for forecasters and communicators alike.

To put it bluntly, counties suck, and we need to stop using them in weather communications.

They're taking care of it. Sort of.

The NWS has taken steps to eliminate the need to issue products on the county level, allowing them to target people affected by individual weather events more accurately and effectively.

1) Grid Forecasts

When you search for your forecast on the NWS website or click your location on the map, you're receiving a point forecast. Instead of issuing forecasts for the entire county, they issue forecasts based on a grid system. Each tile in the grid measures around 2.4 square miles, allowing for a hyper-localized forecast. This is a great improvement over the county-by-county forecasts; some counties are huge, and during active weather events conditions can change dramatically over very short distances.

2) Warning Polygons

Severe thunderstorm, tornado, and flash flood warnings are now issued based on the size and direction a storm is travelling. These storm-based warnings are often referred to as "polygons" by meteorologists. Even though most severe weather only impact a relatively tiny area of a county (on average, tornadoes are only a few hundred feet wide), the NWS used to issue warnings for the entire county. This unnecessary warning coverage led to an increase in false alarms and heavily contributed to the "crying wolf" syndrome.

The polygons allow forecasters to target the people affected by the most severe weather and leave people not at risk for dangerous conditions alone. In theory, anyway, but we'll get to that later.

3) County Splitting

Some WFOs do split larger counties into smaller entities for forecast and warning purposes. The NWS office in Sterling, Virginia, which covers D.C. and Baltimore, recently announced that they're splitting in half some counties in their forecast area in order to better serve residents.

The Washington D.C. area is a great example of an area that sees dramatic changes in weather over short distances. The region's geography creates a microclimate that provides some of the most exciting weather in the country. Bound by mountains to the west and the Chesapeake Bay to the east, the area can see blizzard conditions, destructive ice, and flooding rains at the same time all within a 50 mile east-to-west span.

In response, NWS Sterling has split many of the counties in its warning area between east and west. The western sides of many of these counties often see snow while the eastern half experiences a cold rain. Forecasters have to issue products like winter storm warnings for the worst weather in the county, so folks who live in areas not affected by winter weather are still included in the warning.

County splitting helps target products more effectively, but it's still inherently based on the antiquated county system.

These fixes still have some pretty big problems.

While these are positive steps towards eliminating the use of and providing people with more accurate and targeted weather products, there are still some major technical and bureaucratic limitations.

1) The Emergency Alert System

We've all seen that annoying screen pop up on television or suffered through that ear-piercing screech on the radio. The Emergency Alert System (EAS) has proven an incredible tool to alert people of severe weather even if they're busy listening to the radio in the car or watching the latest episode of their favorite cable trash show.

While the EAS is likely responsible for saving lives, it has some pretty major flaws. The EAS is activated by devices installed in local cable company offices. These devices work by listening to the screeching tone transmitted via NOAA Weather Radio. Those tones are slightly different for each product, and the machine interprets the differences and activates the EAS on television and radio stations accordingly.

When a tornado warning is issued anywhere within Mobile County, for example, the EAS will go off for everyone in Mobile County and for those who receive their network, cable, or radio stations that originate from Mobile.

In other words, it cancels out the usefulness of having polygon warnings that warn only the people who need to be warned.

2) Tornado Sirens

People still rely on tornado sirens for some reason. They're terrible, and the "siren mentality" has and continues to get people killed in tornadoes. Some people who live in rural areas that have tornado sirens will only take shelter when they hear the sirens. They figure if they don't hear the sirens, there's no danger. Nothing could be farther from the truth.

Tornado sirens are only meant to be heard outdoors, for one, and they're wildly unreliable. Some municipalities set them off for any bad thunderstorm—whether or not it has a tornado—and some fail to set theirs off entirely. And most of the time, tornado sirens are activated county-wide even though, again, we have polygon warnings.

People who are indoors should have other ways to receive tornado warnings, period. Blind adherence to tornado sirens needs to end.

We're still using counties!

Slice 'em and dice 'em, draw polygons around storms, and burn all the tornado sirens if you want, but at the end of the forecast period, the NWS is still using counties to produce forecasts and advisories for the public. We need to get over these arbitrary political boundaries (both figurative and literal) and start forecasting the weather for what it is and not what we think it should be based on our constructs of convenience.

Look at these County Warning Areas again:

What weather system fits neatly into those jagged globs of bureaucracy? Try as you might, you can't gerrymander the atmosphere. If nothing else, we need to eliminate—or at the very least rethink—the boundaries that make up a Weather Forecast Office's area of responsibility.

County Warning Areas Breed Inconsistency

Much like Canadians complain that our weather maps stop at the border, as if the weather ceases to exist once one crosses form the U.S. into Canada, weather forecasts issued by each WFO just stop at their borders. Rainfall forecasts, temperature forecasts, snowfall forecasts, tornado warnings—each office is only authorized to issue products for their jurisdiction.

Forecasters have minds of their own. Two forecasters could look at the same model data and the same observations and arrive at two slightly different forecasts based on their biases and previous experience. Most of the time, these slight differences from office to office are hardly noticeable to the general public, but sometimes it creates major problems.

The above map shows forecast low temperatures on Sunday morning. Temperature maps are a great example of large but mostly harmless inconsistency—forecasters at NWS Huntsville believe that lows will be 5-10 degrees warmer in their area than those predicted by all surrounding NWS offices. The effect is especially pronounced during snowfall events, where NWS offices can wildly disagree on amounts (see Binghamton vs. Buffalo last March).

The most dangerous inconsistency, though, comes in the form of tornado warnings. The Vane has covered this issue before, when one particularly ugly example occurred when a tornado-warned storm crossed County Warning Area boundaries three times.

NWS Wilmington believed that the storm had enough rotation to warrant a tornado warning, while the NWS office in Charleston believed it didn't. The result was a ridiculous tornado warning that skipped over an entire county. The storm never produced a tornado, but this type of interoffice disagreement is common during severe weather situations, and ultimately degrades public confidence in the National Weather Service's ability to effectively warn the public of hazards.

It's not just tornadoes and temperatures that prove challenging when it comes to boundaries. Each WFO has its own set of criteria for what kind of winter weather warrants a winter weather advisory or winter storm warning. The amount of snow it takes for counties in Vermont to go under a winter storm warning is much higher than the amount of snow it takes for Greenville, South Carolina to go under a winter storm warning.

For those of us who live on the counties that make up the outer extent of a County Warning Area, it can be extremely frustrating for people who live two miles away to go under a winter storm warning, while your NWS office refuses to say you'll get more than a dusting.

What can we do to fix it?

The first step is the hardest and will face the most pushback from people within the National Weather Service—we need to abolish or radically rethink County Warning Areas. The goal would be to eliminate arbitrary political boundaries and create a seamless, free-flowing weather forecast from one location to the next. Tornado warnings should be issued irrespective of counties or other artificial lines. Snowfall totals shouldn't drop by two feet simply because one forecaster disagrees with another—one of you is wrong, so figure out a fix out for the sake of public confidence.

The hardest part about eliminating County Warning Areas will be to figure out how each individual office will determine its jurisdiction. Blurred lines sound good in theory, but when should NWS New Orleans step on NWS Mobile's toes? What if a situation arises where forecasters at different offices disagree on how much snow will fall or whether or not a storm is severe? These are the issues that need to be worked out, and it can and needs to be done.

It's not just the National Weather Service that needs to rethink its boundaries and the way it issues products. The entire weather community needs to come together to shift away from county-based weather information and refocus our communication on actual impacts as opposed to vague, county-wide impacts.

It will take an overhaul of the Emergency Alert System and a country-wide effort to eliminate tornado sirens from the public psyche. It will take a rebuild of NOAA Weather Radios to activate for storm-based warnings as opposed to entire counties. It will take further efforts to educate the public on how to properly receive and interpret weather information.

Sweeping change won't be easy—it never is—but it must be done in order for everyone in the weather community to more effectively carry out the ultimate mission of keeping the public safe and informed.

[Images, in order: NWS Gaylord via Facebook, author, NWS via NAP, ESRI/NWS, NWS, NWS Sterling, YouTube, author, NWS EDD, author]


You can follow the author on Twitter or send him an email.