My weather app, which gets weather data from The Weather Channel, tells me the weather for NYC. Does all of NYC have the same weather? Does is vary from neighborhood to neighborhood, or block to block?
When I see the weather for NYC on my app, what area does that cover?
Typically it will be the latest hourly report from the nearest weather station to that location.
Usually once or twice an hour, weather stations produce an automated report called a METAR which is a string of data that includes temperature, wind, cloud cover, pressure, recent rainfall etc etc. It is in a standard format that is easily decoded (if you know how to do it) and can be converted automatically to a “latest weather” summary in English.
Sometimes the nearest weather station can be many miles from the location you actually want the weather for. If you search for the weather in several nearby small towns you might find that they all have “identical weather” because they are all using the same weather station data.
The weather channel is mostly just reselling you what you can get just as easily directly from the government with no annoying advertising.
Try http://www.weather.gov/ and in the location box at upper left slowly type a location. You’ll see it give you a roster of matching locations. For NYC there are several; the default appears to be a station in Central Park. There are also stations at each local area airport.
If you visit the page for the local forecast office (New York, NY for NYC) you can often find a roster of the actual weather reporting sites they use. The local sections of the national website are not exactly standardized between all the offices. So some hunting around may get you a map of all their weather stations.
Outfits like weather channel and wunderground MAY have extra sensors in some areas above and beyond what NWS operates. These are actually home weather stations operated by enthusiasts. The good news is that gives more spatially detailed coverage. The bad news is it’s far from well-calibrated coverage.
Right … typically the weather that is reported is the weather at the local airport … but there’s several airports in NYC so unless the app says which one, we wouldn’t know …
“Does is vary from neighborhood to neighborhood, or block to block?”
You bet … sometimes it varies a lot … fog on the beach and sunny inland … rainfall rates can vary from bone dry in one neighborhood and an inch or two in another just because thunderstorms are narrow … out here in the West in the mountains, we have significant differences based only on elevation … rain in the valleys, snow up in the hills …
At this moment, JFK airport is reporting 22ºF and La Guardia is reporting 24ºF … that’s fairly close so if your app is reporting from one of these locations you should have a pretty good idea what the temperature is out your own front door, although it might not be exactly these reports … the only time to worry about a large difference over a short distance is if there’s a weather front in the area, sometimes these differences can be extreme …
As I mentioned above … rainfall can be quite varied … it’s entirely possible for La Guardia to get a couple three inches over a few hours and JFK get none … another example is Tropical Storm Harvey’s rainfall in Houston, this was what “looked like” an atmospheric river event, Houston saw all-time record rainfall but just twenty thirty miles either side didn’t get near as much … they got a lot, but nothing close to what we saw in Houston herself …
Bottom line … if you want to know the temperature at your house, buy a thermometer … or see if one of your neighbors has one that is posting to the internet at sites like what the Weather Underground has set up … “hyper-local” weather indeed … and, no, I have no idea why a terrorist organization from the 1960’s is operating weather stations today …
Bottom line, there are observations made hourly at spots all around the city and all around the country. Depending on which parameter we’re talking about these observations involve a mixture of readings taken at one moment, readings averaged over a few minutes, and readings aggregated over the full hour. But in each case they’re talking about one single spot on the ground, usually about 5 feet across.
There are also computer-generated forecasts at various spatial and temporal densities run every few hours. By different computer models. Which are then combined by a mix of computers and human skill/intuition, into numerical and prose statements about expected weather over a span of many miles and many hours.
All of this stuff is “weather reports”
Maybe with the info given so far in the thread the OP can come back and ask more refined questions about whatever he’s really interested in.
The NOAA webpage listed above has a zoomable, scrollable map. You can check out different neighborhoods covered by the same office for localized temperature and precipitation. You do get different weather as a result of the lay of the land, such as tree cover blocking thunderstorms, water moderating temperatures, and the like.
Yeah, my weather underground app says my local data comes from “Dave’s Yard”.
The traditional model of weather observation is as Colophon describes it: an hourly observation formatted in a machine-readable (but also human-readable, if you know how to decode) format. The weather observation point is usually an aerodrome, since the first major use case for current weather observation was aviation.
The observation frequency can be increased to several per hour if conditions are appropriate. In other words, special reporting criteria, necessitated by unusual and potentially hazardous conditions, like developing severe weather.
Still, a normal “nothing interesting going on” obs is taken about 5-10 minutes before the top of the hour, at a fixed location on an airfield, so it can be up to 50 minutes old and most certainly in a different neighborhood than the average user.
Lately, weather service providers are beginning to incorporate mesonet observations. “Mesonet” is a portmanteau of “mesoscale network”. “Mesoscale” is a descriptor of the time and spatial scope of weather operation that indicates that the phenomena being observed or forecast are in an arena measuring hundreds of miles and minutes, rather than continental distances and hourly (the traditional operating scale). As such, mesonets access local official and unofficial weather observation systems, such as registered personal weather stations, to greatly increase observation density and currency.
Providers like Weather Underground incorporate mesonet observations and the WUnderground app on my phone typically picks an observation that originates within a few blocks of where I am according to GPS, and usually just a few minutes old.
Sorry for the wall’o’text. This was my line of work for decades (not forecasting, but design and development of data and comm systems supporting weather operations) so I have a hard time hiding a lot of the complexity.
I just checked … the NWS issues “spot forecasts” using latitude and longitude to 14 decimal places … that’s about one square micrometer … they used to issue forecast to 20 decimal places which is about every 8 molecules of SiO[sub]2[/sub] … you’d think they were compensating for something …
[noparse]National Weather Service
I know you’re kidding …
Never confuse precision with accuracy. Some web dev didn’t bother to filter his coordinates to appropriate sigfigs. Idjits. My comment about 5 feet was the typical footprint of the typical instrument package. Where on Earth it’s sitting and to what degree to we care about that location precision is a separate question.
As to spot forecasts, how many micrometers can you move from a given point before you get a different output? And can you actually walk that far in micrometer increments before the output changes due to time, not due to space? What’s the spatial resolution of their geoid and terrain models?
I provided the link … it really does say “longitude = -100.41174975585938, latitude = 41.5548086239053” … my guess is this extraordinary accuracy hides all their bogus forecasts …
If we click the link, we find the area covered is more like a couple miles squared … and even that seems a tiny unit volume for the computational models … the NWS only has 6 hours to create their forecast packages and I’m not sure they can complete one model run in that short of a time with such a small volume … it’s surprisingly difficult to get accurate measurements in the vertical, and this data is critical, here in The West our radiosonde sites can be a hundred miles apart … but as long as they’re issuing forecasts for each square micrometer, no one notices … [giggle] …
Can you tell me where I can locate the list of reporting sites they use?
Basically, Yes.
The main weather pattern will be true for a fairly sizable area. There will be minor variation in things like temperature, but it will only be a few degrees plus or minus from the average that they give. Usually, it’s raining over most of the area. When it’s not, they usually say something like “scattered showers …” When temperature is changing significantly, they will say something like “a cold front is approaching from the northwest, and temperatures are expected to drop 10-20 degrees as it passes”. Here in Minneapolis-St. Paul, they often say something like “3-4 inches of snow accumulation is expected overnight in the western half of the Metro area, but only 1-2 inches in the remainder”.
You’ve never been to the Bay Area, I take it.
Before we moved here, all the places that I have lived followed the pattern you described. But here we have many microclimates, thanks to the influence of the Bay and the mountains surrounding us. I live maybe 40 miles from San Francisco, and weather there and weather here have little to do with each other. In fact temperatures in SF and south of it on the peninsula are also quite different. Local weather forecasts give at least three readings, and sometimes more.
Where I live you seldom need air conditioning - at least not until recently. 20 miles up 680 in Pleasanton it is a must.
So we’re an exception.
In New York where I grew up only one weather forecast worked for the entire region, though people understood it might be cooler at the beach.
The latitude and longitude are just used to calculate what square you’re in on their forecast grid. They aren’t issuing different forecasts for each micrometer. You put in that you want the forecast for zip code 01234 or whatever, so it takes the coordinates of the center of that zip code and uses them to decide which square’s forecast you get.
“The coldest winter I’ve ever spent was summer in San Francisco” – Mark Twain
I lived many years up by the Skyline just across the border into Daly City … it was routine to have dense fog at my house but downtown SF reporting sunny and 78ºF … that was just 6 miles away …
It may not exist. I gave you a head start and invited you to go look for it.
What are you actually trying to learn or understand? You seem to have some goal in mind, but you haven’t shared it with us. I suspect there’s confusion in your goal and if you tell us what’s really on your mind maybe we can figure out how to get you what you really want, not what you now mistakenly think you want.
See also t-bonham above. Simplifying just a bit … Weather is *not *what’s happening at your house this instant minute. Barring special areas like SF, weather is what’s happening to the most of the county you live in over the last hour or next few hours or so. It only gets that precise.
And Colorado Springs. Through personal observation, I know of 4 different microclimates within a 7-minute drive from my home. The differences can sometimes be quite dramatic.
I want to know where the temperature I get when I open my app comes from. Where is the data being reported from? I get that it might be from one of the airports in the city or from central park, but how do I know which reporting site was used specifically?
I don’t know what app you use but one that I have, Weatherbug has a network of basic stations (temp, baro, wind) at elementary schools in my city. It uses the one closest to my location - about a mile.