Monday, January 31, 2011

Forecast Wrap-Up

My apologies about the absence of a final post on Saturday. A few factors played into it: I was not at work in order to create my graphics and I actually did not have any Internet at my house on Saturday. Excuses, excuses I know. I may be looking for excuses for my forecast on Sunday as well! Things really didn't pan out the way I thought. In my last post I had determined with some confidence that Sunday would start mostly cloudy and eventually the clouds would part in the afternoon, remaining dry all day. But that didn't happen. One reason (excuse?) why...the models seemed to delay the rain!

In the few days leading up to Sunday, rain was forecasted for late Friday night and into Saturday. That didn't happen! Saturday was mostly dry But what did happen was a mostly dry, mostly cloudy Sunday with some rain that moved in late afternoon/early evening. Models tend lag or even accelerate some times and that can throw a wrench into the process of forecasting. The only way to maneuver around that is to know how certain models handle certain weather systems. I haven't been looking at the models long enough to pick up on the their tendencies. That will come with time. So really my conditions forecast was not all that far off except that rain ruined the accuracy of the forecast. I will be one of my toughest critics and give myself an X on the conditions but only by 0.03" (recorded amount of rain at the Portland International Airport).

My final forecast had a temperature high of 48 degrees. My forecast low was 38. Now, in order to give myself full marks for an accurate temperature forecast, I'd like to be with in +/- 3 degrees of the actual observed temperature. Sunday saw a high of (drum roll, please)...50 degrees! That high temperature occurred sometime between 2-3PM, typical for this time of year when our sun angle is rather low. During the summer we get our warmest temperatures of the day between 4-5PM. So my high temperature forecast was a success, missed it by 2 degrees. For the low, Sunday got down to...39 degrees. Missed it by one degree! Not to shabby.

If you would have asked me which I had more confidence in, my conditions forecast or my temperature forecast, I would have gone with the conditions. At least I got something right! I hope that this process allowed you to look into how I do my job and all the moving parts associated with it. I know weather forecasters, ones on TV especially, get a bad wrap. We are forecasting for an environment that is always changing. Portland and the Pacific Northwest as a whole is challenging to forecast fore given that most of our storm systems come off the Pacific ocean where current observations just aren't as abundant as if we lived in Kansas City. The Pacific ocean is dotted with some weather buoys but those don't help us nearly as much as observations from other cities. Places all across the central and eastern U.S. get to see what is approaching them and how an air mass headed their direction has been impacting the cities to their west. Also, the amount of model information out there is large. Many models and charts are at the forecasters fingertips and it can be a challenge to select which one is the correct one. All this information must be taken into account. It is definitely a challenge but someone has got to do it. I am glad I can!

Friday, January 28, 2011

Day 5

Alright, we are closing in on Sunday. With only two days to go we have reached the time where forecasts are now becoming much more accurate. On average, forecasting accuracy drops significantly when you get 4-5 days out, yet several weather services produce 10-16 day forecasts! There are no excuses at this point with close to 48 hours to go! Today, I have completely taken out the chance of rain for Sunday leaving us with a partly cloudy afternoon. We should get some rain today and tomorrow though. Sunday begins another stretch of precipitation free conditions that we have been enjoying for the last week or so. After analyzing today's maps I will include a bit on how I forecast temperatures.

Most of the action remains to our south on today's models. The development of a second low near Alaska has also helped drag moisture further to our north that will follow that establishing ridge that shows up quite well on the 500mb map. Look at the big high pressure region to the north of Montana. That is really cold arctic air. Temperatures look to be in the single digits over the weekend with lows in the negatives for Montana, Wyoming, North and South Dakota (I just looked at the forecast temp for Lemon, SD...the HIGH temp is 3 for Sunday, ouch!).

We see the same thing happening at the 850mb level that we see close to the surface. The further up in the atmosphere you see features, the "deeper" they are considered. In this case, the lows are fairly deep, although it is not uncommon to see features all the way up through the 300-200mb level, like hurricanes. We see little hope for some cold air advection, which would stabilize the atmosphere. Nonetheless, we still are looking dry.

That big ridge is still lingering out near the coast. It acts as a wall that forces energy (PVA) up and over the ridge. This ridge also impacts our temperatures at the surface. Under them, locations can expect warmer than average temperatures. We will take a look at how this impacts our temperatures in a bit with the numerical forecast.

The weak jetstreak that we are dealing with looks more harmless today. It is pushing south, we call this "digging". Notice the little dip in the overall flow to our south. That is a small trough, caused by that shortwave energy. As long as the jetstreak stays on the LEFT side of the trough axis, we can expect that energy to "dig" and deepen that trough. That should send the shortwave even further south. In fact, it will impact southern California this weekend. For us, we will be enjoying dry and partly cloudy conditions.

I haven't really spoken much about temperatures for Sunday yet. There are a couple ways I go about forecasting temperatures. As I noted in the into, forecast accuracy declines when you look to predict 4-5 days out. I typically do not use these maps for high and low forecasting. It is possible, however. A lot of forecasters use the air temperature at the 850mb (in degrees Celsius) that corresponds with a temperature at the surface (in degrees Fahrenheit). For Sunday's forecast, I have used two sources. One is the 2m surface temperature model that the University of Washington provides. It forecasts 84 hours out and uses 3 hour increments. This is about as high-res as we can get in forecasting for our area.
Forecasting a high temperature is easy enough using this method, assuming you have confidence in this model. Over a year's worth of forecasting has lead me to notice some trends (flaws, in my opinion) within the models. Every model is great at handling some aspects and struggle with others. As the map shows, we are right at the threshold of 48 degrees. I am holding onto hope that this model keeps us right there since that was my initial forecast temp!
Another method used to forecast is model output stats, or MOS. MOS is great for the Pacific NW, Oregon and Willamette Valley forecasting. Why? MOS has a strength in handling local effects really well. Local effects include the Pacific Ocean, Coast and Cascade mountain ranges and the Willamette Valley. All these things are taken into consideration. Here is what MOS looks like:
A lot of categories there! Some include dewpoint temperature, cloud cover, probability of precipitation for 6 hours, probability for 12 hours, type of precip, ect. What we are concerned with is temperature. You'll have to forgive the skewness of the graph but the first few days mostly line up correctly. Look at the row labeled "X/N". That is the nighttime/daily max temperature. Go to Sunday, 72 hours out (this was yesterday's model, they re-run at 0z, which is 4p.m. our time). MOS is expecting a high of...48! A great way to re-enforce a forecast! This particular MOS has several perturbations on it, about 20. The above is just one of those. I averaged all the perturbations and the temperature came out to about 47.5 degrees. I'm sticking with 48, why change based on half a degree. You got to own your forecast!!

Thursday, January 27, 2011

Day 4- Riding the Ridge

We are at the proverbial "halfway point" to Sunday. With each day, we have seen an increase chance in precipitation and not much change in the way of high/low temperature. We are on the cusp of getting much more data starting tomorrow. That is when multiple models start their Sunday runs. For the past 4 days I have been using extended GFS (Global Forecast System) models to forecast. GFS, NAM (North American Model) and more GFSx models will be available for use. Also, the University of Washington has a very good model system set up that uses NAM and WRF (Weather Research and Forecasting) model that specifically focus on the Pacific Northwest. That will help as well. So with more data means more information to sift through. This is where the fun really begins. But as for today's update:

Yesterday we saw a nice batch of moisture being pushed our way with the help from high pressure to our south and low pressure to our north. Today that moisture has been pushed back into the flow of the low pressure system. That moves the precipitation to the north and removes a good chance of rain for the majority of the day Sunday. Due to this, I have removed the showers for most of Sunday.
By the way, check out the Pacific weather charts as of today. You can see that area of low pressure that will eventually make its way into the Gulf of Alaska. It's sitting out in the middle of the Pacific. Pretty neat that we can see this coming from so far away!

I am adding a new map to the discussion today. This is the 1000mb map and is essentially a surface map. Many locations around the state would call 1000mb "ground level". What I am looking for in this map will be re-enforced all the way up to the 500mb map today, more on this in a bit.
I have highlighted a few wind vectors. Wind vectors on this map indicate direction and wind speed. The longer the arrow, the faster the winds. Notice how the winds around Oregon spread apart from one another. This indicates that wind at the 1000mb level are diverging away. This is cause when air sinks from the upper levels. When we have air sinking, once it hits the ground the only thing it can do is spread out. Sinking air also denotes stable conditions and high pressure. This is another sign that our rain chance for Sunday is diminishing at this point.


High pressure off the California coast is still the dominant feature at 850mb. The low up to the north weakened on the maps from yesterday. There appears to be little to no warm air advection on today's map. That favors more stable conditions.
That pesky shortwave makes a return in today's map but it doesn't impact our forecast. What does impact the forecast is the ridge that has been re-established. This ridge is just off the coast. I would not be surprised to see this ridge move a bit more between now and 4pm Sunday. This ridge is re-enforced at the 1000mb level. Remember the divergence of wind on that map? It is a result of this ridging.


The jet streak we saw yesterday that was set to help with the precipitation has now left. It has sagged to the south along with that shortwave energy.
All in all, day 4 has brought an improvement in the forecast. I would say the only threat of rain on Sunday will be limited to the early morning hours. The first four posts have been strictly conditions updates. Tomorrow I will get into some number forecasting as various models and text models will begin to include Sunday. As of today, my initial numbers have held up pretty well!

Wednesday, January 26, 2011

Day 3

On to day three of the forecast test. Still a few days out from getting better data for the 30th. Currently I am using just one model (something called "model-riding") but we have to get within 72 hours in order to get a look at multiple models. We are still looking at an ever-changing situation for Sunday. I am see another increase in chance for precipitation on Sunday. It doesn't look like a real big rain event, just some scattered light showers. Let's get into what has changed:

Yesterday the surface model was forecasting some moisture from the south moving into our area. Today that threat has diminished. Off the California-Oregon border, high pressure has built northward. That will interact with a developing low pressure system establishing itself in the Gulf of Alaska. In the Northern Hemisphere, high pressure is associated with a clockwise spinning rotation. Low pressure in the Northern Hemisphere is associated with a counter-clockwise rotation. Look how these two rotations interact with each other. They wind up converging right in our region. That will help send a little disturbance our way.


The 850mb map re-enforces what we see at the surface. The higher pressure that built northward is surrounded by warmer temperatures. The low pressure that has developed contains cooler temperatures. With the rotation of each system kept in mind, we see that there will be a convergence of warm and cooler air. This is what's called Warm Air Advection, areas where warm air pushes into an area of cool air. The WAA isn't strong but nonetheless it helps that we have a little to enhance uplift and some unstable conditions.

The first few models that we looked had a prominent ridge over our heads. In today's model, that ridge has flattened out a bit. Remember that shortwave that I talked about the past two days?? It has disappeared from our region. It sagged even further south and merged with a trough that was almost off the map the first few runs. The fall out of this shortwave means that we won't be getting the moisture from the south. That energy falls to Baja California. So where will we get the energy for the moisture??


Our last map reveals the new source of energy. First off, this map re-enforces the shortwave that moved south. Look at the trough that sags all the way down to Baja California. The shortwave followed this flow. Back in our region, there is nothing that impressive to look at. However, there is a small jet streak to our north. This streak is located over northern Washington and Idaho. The rule of thumb with a jet streak is this: look at the jet streak from behind it as it flows west and split the streak into quadrants (as I have done to the image below). There are two quadrants to take note when looking for weather action. The front left quadrant (known as the left exit quadrant) and back right quadrant (right entrance quadrant). These are areas of PVA, or positive vorticity advection. PVA is associated with areas of uplift, which causes air to rise, cool and condense out often leading to precipitation. Well guess where Portland is in relation to the jet streak. In the right entrance quadrant! That is right where we want to be when looking for some precipitation action. Again, our jet streak is fairly weak. Look at the jet streak out in the Pacific, that is a strong streak associated with that low pressure system up in Alaska.

As you can see, each day brings change. The atmosphere is always changing and it is easy to see why there are so many revisions to a forecast, especially during the winter months! Nonetheless, not many changes have been made to the initial forecast. It helps that this doesn't appear to be a big winter event, too! Back again tomorrow.

Tuesday, January 25, 2011

Day 2

Things have changed slightly for our forecast day, Sunday the 30th. The temperatures remain the same. In fact, the evening forecast kept the same numbers last night. The change in the forecast comes in an increased chance for rain. Let's take a look at what changed on the maps:

Let's start from the ground up.

Surface Map
Here is the latest 0Z update for the surface. We see the batch of precipitation spreading northward. Yesterday, this moisture was limited to southern Oregon. Now it has pushed its way a bit further north into our region. I have added some edits in order to highlight the areas that have undergone change.

850mb level (about 5,000 ft)

Using this map, I am looking to diagnose where areas of warm air invade areas of cool air and vice versa. Areas where warmer air is over-taking cooler air indicate instability and aid precipitation processes. Areas where cooler air is moving in will indicate stable conditions. In our situation, we aren't seeing any warm air coming in, nor is a significant amount of cool air coming in. All in all, today's map isn't all that impressive.

500mb level (18,000 ft)
As I mentioned in yesterday's post, the 500mb level is the steering mechanism for a majority of our weather systems. Yesterday, we saw a good ridge over our heads but to our south we saw a shortwave, cutoff low. In today's map, that cutoff low has weakened some. The energy is still there. The region is still centered in southern Oregon and has plenty of positive vorticity associated with it. This is a large area of uplift, which is a major part in aiding precipitation. This particular area still looks strong enough to produce some good uplift in our area, which is why I've upped the rain chance a bit. But the low was cutoff yesterday but not today. We look to the 300mb level to see why.

300mb level (30,000 ft)

This map shows the location of the jet stream as well as streaks within the jet flow. Yesterday, we saw a jet streak that was beginning to dig south. Today, it has flattened out a bit but increased in intensity. That is helping push the energy up our way. The southerly flow yesterday was a sign that the jet may dig south, allowing the cutoff low at the 500mb level to go south as well. Now that the jet has flattened out, that cutoff low will now begin to flow back up towards the north. The cutoff low itself has weakened but the jet has increased in intensity. This will be another aid to precipitation as an increase in the jet means an increase in uplift around the jet streak.

So the glaring change from yesterday to today was the intensity and location of that shortwave kink in the middle of the atmosphere. It will be interesting to watch the shortwave and see if any cold or warm air moves into the region. Also, the closer we get to our forecast day, we will have access to more model runs that will allow some comparison. I'll be back tomorrow with another look at the forecast!

Monday, January 24, 2011

7-Day Acurracy

Mark Nelsen, known around work as "The Chief", grades his forecast accuracy each week. Daily high/low temperatures and sky/precip conditions all go into his grading scale. He churns out a pretty high accuracy rate. His report card got me accurate can we forecast seven days out? A lot of people think that meteorologists can't even forecast 1 day out so how accurate could we be seven days ahead?

In all reality, forecasting is becoming much more accurate overall. When it comes to you certain aspects of forecasting though, things can be more challenging. Take for instance winter weather. Everyone gets upset at the weatherman when says it is going to snow and then it never does. Why does it always seem that way? First, we live in a geographical area that doesn't often produce snow where we live. Being in a valley, temperatures really have to sink down to the floor in order for snow to survive. The Pacific Ocean also acts as a moderator and will warm an otherwise cold air mass that could produce snow. Water has a high heat capacity which means that it take much more energy to warm the water. Because of this, the winter time is actually when water temperatures in the Pacific are WARMEST. Another reason why forecasting winter weather is so tough? We just do not know what will happen until right before it actually happens. So many dynamics go into winter weather and it only takes the difference of a few degrees that would give us snow, rain or even freezing rain. So next time you see snow in the forecast, just be patient and keep in mind that things can (and probably will) change.

Now onto my little experiment. I've decided to track my forecast for next Sunday, the 30th, for the next seven days. We will see how much it will change and see how accurate it is. Let's look at day 1:

Ok, so seven days out I see us staying dry. Things cloud up after a decent stretch of sunny conditions. The average for January 30th is 47 degrees. I am shooting for 48 degrees. This many days out, climatology can be a decent way to forecast a high and low temperature. A lot of hi-resolution weather models do not go out 7 days so my confidence on this as of today isn't all that high. But I am, as of this moment, fairly certain we will stay dry Sunday. Here is why:

Looking at the weather charts, a good place to start is the 500mb. This is about half way up the atmosphere and gives a good estimate of where weather systems will go.

This is the 0Z (+8 hours) 500 mb model run, valid Mon. Jan 31. This is our 4 p.m. forecast for Sunday. Two things to point out. Notice the white contour lines over Oregon and Washington. They form a ridge in the overall flow. A ridge indicates a region of high pressure and often results in fair conditions. That is a pretty well-established ridge over our heads. But look to the south, right around the Oregon-California border. That is a shortwave disturbance, kind of a kink in the overall flow. Shortwaves often result in unstable weather. So we have a ridge overhead, but a shortwave of energy to the south. What will this lead to?

The map on the left is a relative humidity chart and the map on the right is what is expected at the surface. Again, these maps are all forecasting for next Sunday. On the relative humidity map, treat the lighter areas as clouds. Now look at the that bulls-eye area at the Oregon-California border. Showing a pretty good area of high relative humidity, meaning the air is close to saturation and is a stronger area of potential rain. Now turn to the surface map. There is some precipitation near the border. But in our area, we look dry. As these models update we can get a better idea of the flow and where this shortwave may actually land. It is possible this area moves a bit further north and impacts us. But as of right now, all I see are some clouds from this shortwave that could filter into our area, thus the cloudy yet dry forecast.

What do other forecasts look like for the 30th? The Weather Channel has a high of 48, low of 42 with a few showers. The National Weather Service also has a forecast high of 48 under mostly cloudy skies. So at least other people are seeing what I am seeing. I'll be back with an update on the forecast tomorrow.

Thursday, January 20, 2011

Putting A Cap on the Season

I am slowly coming to terms with seeing my favorite sports team lose in the championship game with 2 seconds to go. What a ride, nonetheless. A great season ended with a great game regardless of the outcome. The National Championship game was a classic, one of the best bowl games of the season. But in my mind, it wasn't the game of the bowl season. I'll talk about which game gets that nod in a bit. First, let's re-visit my bowl picks.

Ultimately wound up with a 20-15 record. Preeeeeety average but I'll take an above .500 season. There were a few surprises, as there usually is. I was surprised how Stanford romped, how Alabama rolled and how Oklahoma didn't get 60 on UCONN. Nebraska choked, Miami fell flat and Ohio State tried but Ryan Mallett tried harder to lose and the Bucks finally beat an SEC team.

Things that went according to plan: TCU and Wisconsin played smash-mouth football, the National Championship game was low-scoring (it always seems when all the talk is about the offense, the defenses step up), Arizona continued it's horrific 2nd half of the season. Michigan was shown by Mississippi State that it didn't belong in a New Year's Day bowl.

The bowl with the most deserving winner: Army in the Military Bowl
The bowl with the most ironic loser: Hawaii in the Hawaii Bowl
Most boring bowl game: UCF over Georgia (1 TD, 3 FG's)
Most exciting bowl game: North Carolina over Tennessee- UNC shows that committing a penalty isn't always a bad thing. Last second kick sends game into OT and the Tar Heels steal one in double OT.
Most to gain from bowl win: Stanford. Could have been ever bigger had Harbaugh stuck around. Nonetheless, Stanford will be in the talk for a Pac-12 championship next year and you will hear about them in the National Championship talk as well.
Most to lose from bowl loss: Miami. They were the only Florida school to lose. Schools like South Florida and UCF will definitely benefit from their bowl wins AND Miami's loss.

Too early to look forward to next year. Still digesting the past season, it was one to cherish. Spring ball is not too far off and that's when things will get serious.