Seasonality: Backstage

Seasonality is a family of apps dedicated to tracking, analyzing and understanding the weather. GetSeasonality.com

[getseasonality]

Sunsetting Seasonality

All good things eventually come to an end and the time has come for Seasonality apps to sail into the sunset. When I started working on Seasonality, there were only a handful of serious weather apps for macOS (or OS X, as it was called at the time). Now, almost 21 years later, the Weather category on the App Store is filled with thousands—even tens of thousands—of high quality options.

Gaucho Software has always been a one-person shop. Over the years, I have contracted with a handful of people for graphics and app design work, and I greatly appreciate all of their contributions. However, I've always been the sole developer on the project. In the early years, Seasonality provided my full-time income, which allowed me to focus entirely on improving the apps. As time went on, I began taking on contract work to both supplement my income as well as add some variety to the code I was writing each day. This reduced the amount of time I could spend working on Seasonality but still allowed me to maintain it at a reasonable level and add new features from time to time. A few years ago, I found myself looking for a change and decided to take a full time job as an iOS developer, working on an app to help medical professionals. I love working full time as part of a larger team and contributing to improvements in healthcare, but this leaves me with no meaningful time to continue supporting Seasonality.

For a while, I thought I could just let the apps coast. Sales continued to trickle in, and they have been enough to cover the server costs required to keep Seasonality's weather data available. However, as new OS releases come out, things inevitably begin breaking in the Seasonality apps, and I don't feel comfortable charging people for software that no longer meets my own expectations.

So with a heavy heart, I have decided to let go of Seasonality. I will continue to keep the weather data servers online, but I won't be actively maintaining the apps. That means something different for each version of Seasonality. The plans for each are outlined below.

Seasonality TV The tvOS version of Seasonality never really took off, and has been particularly difficult to keep up-to-date because it's codebase differs significantly from the other versions of Seasonality. For this reason, Seasonality TV will be removed from the App Store sometime soon.

Seasonality Pro This app depends heavily on model data hosted by NCEP, which is outside my control. Seasonality Pro will continue to be available as a free initial download, as it is today. In addition, a future update will make all VIP subscriptions free. If you are currently a VIP subscriber, your subscription will not renew when it expires and you will continue to have access to the same features for free. The app will remain available for the time being, assuming the underlying weather data is online.

Seasonality Go I use this app on my iPhone every day, which makes this decision especially difficult. I'll be removing it from sale soon, but Seasonality Go may still receive occasional updates. If something major breaks and I'm able to find the time to fix it, I will likely release an update and briefly return the app to the App Store so existing users can update. After that, the app will be removed from sale again to prevent new purchases.

Seasonality Core Since this is an app for macOS, I have some additional flexibility because it can be distributed outside the App Store. I'm currently working on a bug fix update to address some issues in macOS 26. I will submit the update as the final App Store release, and then remove it from sale a few weeks later. I then plan to make a free version available via getseasonality.com, giving me the option to address major issues in the future if time permits.

While some versions of Seasonality will continue to function (and existing installations should continue to work) for the time being, I won't be able to provide user support going forward. Unfortunately, I haven't been able to offer much email support for quite some time, which is part of what led to this decision to begin with. Server support will continue for now, but it may be discontinued at any time.

I would like to thank everyone who has purchased any of the Seasonality apps over the past two decades, especially those who took the time to write in with feature requests or bug reports. Thank you to the government and educational meteorological organizations as well that have made their data freely available. Thanks to Chris, FJ, Jochen, and all the others who contributed to the design and user experience of Seasonality apps. Finally, a heartfelt thank you to my wife and family for supporting me in Seasonality's development for so many years. I truly couldn't have done it without you.

With that I'll say farewell, until we meet again.

[getseasonality]

Designing Professional Meteorology Tools

I gave a talk at the Annual Meeting for the American Meteorological Society in Seattle back in January.  The presentation was recorded and a video was recently posted.  During the 15 minute presentation I discuss a lot about my thought process of designing Seasonality Pro and during the second half of the presentation I give a sneak peak on some features I’m working on for the next update.  Check it out…

https://ams.confex.com/ams/97Annual/recordingredirect.cgi/id/38052?entry_password=104527&uniqueid=Paper313398

[getseasonality]

Observations update (fixed)

The Seasonality Go 4.0 update was approved by Apple and everything should be fixed.  To summarize, upgrade to Seasonality Go 4.0 and Seasonality Core 2.5 in order to fix the issue of missing graph and current conditions data.  Thanks again for everyone’s patience while I resolved these issues.

[getseasonality]

Observations update 3

A Seasonality Go update is coming very soon.  This past Saturday I submitted an update to Apple, and I’m just waiting for it to pass through the App Store review process.  I have yet to receive a notification that they have started reviewing it yet, but am optimistic they will start looking at it shortly.

[getseasonality]

Observations update 2

This week has been productive on the changes needed for Seasonality Go to use the new data source.  The data is now loading in the app, but I’m still working on some bugs in the graphing code to make sure this is a quality release.  I’m hoping to start a beta-test as soon as the bugs have been fixed and everything is running smoothly.  Assuming the beta test goes well, I’ll be submitting the update to Apple soon.

Again, thank you for your patience.

[getseasonality]

Observations update

A lot of progress has been made since the data source for weather observations went offline a little over a week ago. 

Like I mentioned in the last post, the new data source is hosted on Seasonality’s weather servers.  This data source has been used for a couple of years by Seasonality TV for the Apple TV, but I needed to fix some performance issues before opening it up to the much larger audience of Mac and iOS Seasonality users.  I believe the servers are in good shape for the update now.  The added benefit of this new source is that it can provide a lot more data to Seasonality.  Now you’ll be able to download the last 7 days worth of data instead of just 4.  I’m hoping to increase this further in the future.

Seasonality Core 2.5 has been updated to use the new data source.  You should see an update in the App Store sometime today (it’s going live as I write this).  This update supports macOS 10.7 or later.

Next I will be working on Seasonality Go and hope to have an update ready sometime in the next week.  The Seasonality Go update will include some other improvements I’ve been working on for awhile as well, so I look forward to finishing it up.

[getseasonality]

Observations data source offline

Late last week, the data source that Seasonality Core and Seasonality Go uses for current weather observations and past graph data went offline.  This source on NOAA’s Aviation Weather Center site has been online and providing Seasonality with high quality data for most of the last decade, so it’s pretty incredible that the service has been so reliable and I haven’t had any problems before now.

I’m using the opportunity to switch over to a new data source I’ve been testing with Seasonality TV for quite some time.  This source is hosted on Seasonality’s own weather servers, so I’ll have more control over fixing any problems that might arise like this in the future.

Hoping to have an update for Seasonality Core sometime this week, and Seasonality Go next week.  Sorry for any inconvenience this has caused for Seasonality users. 

[getseasonality]

Here's a sneak peak at the new maps in Seasonality Core 3.0. There are several notable changes in the image above. Possibly the most obvious change is the new base map. Previous versions of Seasonality Core used NASA Blue Marble satellite images. They were beautiful images, changing for every month of the year to see the differences in foliage and snow cover. However, they didn't have the high resolution of street-level maps. Seasonality Core 3.0 will be switching to Apple's MapKit maps. This gives street-level detail along with the more familiar Mercator map projection. The other obvious update is in the radar resolution. With higher resolution base maps, expectations are higher for the map overlays as well. While the current radar imagery is alright in the broad sense, I have been looking for something more for quite some time. You will still see the same composited images when zoomed further out on the map. However, when you zoom in enough Seasonality Core will automatically switch to the new Level III images. These are drawn in vector format, which looks very sharp no matter how far you zoom in. Only the radar station closest to the center of the map will draw. This keeps neighboring stations from colliding with each other. For radar stations outside the U.S. that don't provide Level III data, the highest resolution bitmap image will be used instead. On the top of the map, you'll notice a new set of controls. On the left side, you have a configuration button to select what you want to see and how you want it to be displayed. There are some new options in this menu that I'll talk about in the future. In the center is the date display showing the time of the current frame as well as a color legend. On the right are new frame controls. The app will animate 10 frames of images, and now you can control which frame you want to see in addition to animating the whole set. Overall, the map is much improved in the new app. I don't have an estimated release date for Seasonality Core 3.0 yet. New features are still being added, and then the app will need to be beta tested. Stay tuned for more details here as time goes on.

[getseasonality]

The Seasonality Observations Server

With the recent introduction of Seasonality TV for the new Apple TV, I decided it would be a good opportunity to test the waters on a new weather observations server I’ve been working on for quite some time.  

Weather data is big, and it takes clever coding and servers with fast processors and storage to manage it.  Weather observations are a perfect example.  The new Seasonality observations server keeps track of weather conditions (temperature, wind speed, etc) for over 50,000 weather stations around the globe.  Whenever those weather stations report a new observation, the server will retrieve that observation and store it in a database.  Every time someone opens Seasonality TV, the app will download the current conditions from the server and show them to the user.

All of this has to happen very quickly.  New observations are downloaded to the server every 15 minutes, and should only take a minute or two to parse and insert into the database.  When a Seasonality app wants to fetch current conditions, the server should be able to search the database and respond with the observations within a fraction of a second.

It was quite a feat to achieve these speeds.  I developed one system to store observations that just wasn’t fast enough.  So then I had to scrap it and go back to the drawing board to design the current system.  The new system is a lot faster and should be able to scale to support more weather stations in the future.

I have imported about 1.5 years of past data into the Seasonality observations server.  New data is being added constantly, of course.  To give you a sense of scale, there are currently over 800,000,000 observations in the database for the 50,000+ locations.  At this rate, the server will be storing more than a billion observations within the next 3-4 months.

Once the server has been thoroughly tested, I’ll bring it to Seasonality Core and Seasonality Go.  This will offer an extra 40,000 weather stations than are currently available in those apps, which is great for users in rural areas that might not have a weather station nearby.  

If you have a personal weather station, you can upload your data to the Citizen Weather Observer Program, and your weather station will be added to the Seasonality observations server as well.  This will give you a way to view your backyard weather directly in Seasonality (once Seasonality Core and Seasonality Go start using the new server).  The CWOP website has more information on how you can participate.

[getseasonality]

Seasonality Pro for Surf Conditions

I grew up bodyboarding on the Central Coast of California.  Pretty much every weekend, you could find me out with some friends at Pismo Beach.  So when looking over the models I had access to for Seasonality Pro, the WAVE model really caught my eye.  

There aren't many variables in the WAVE model output, but there's enough to be pretty useful.  Primary and secondary swell directions, heights, and periods are all available, as well as surface wind (which can easily come from any of the other weather models).  

To make things simple, I've created a preset to show a blended layer with the primary swell height and a separate arrow layer to show the primary swell direction.  Should be a nice map for bodyboarders and surfers to check out before getting wet.  

[getseasonality]

An Unexpected Performance Test

I’ve been creating some built-in map presets for Seasonality Pro, to make it easy to look at the most commonly used meteorological charts.  They are looking pretty cool so far:

Each time I create a new one, I test it by adding a new map using the preset.  Then I check to see if it looks right and make any necessary adjustments before trying it out again.  I repeat this process until the preset looks perfect.

While working on this process, I haven’t yet finished adding a delete map button though.  So I have some good news: Seasonality Pro can handle 60+ custom maps without skipping a performance beat. :-)

Now to work on that delete button…

[getseasonality]

Winds of Change: Making a new Seasonality Go Icon

The following is a guest post by Chris Emery, the owner of Cwench. Chris worked on the new icon and other interface images for Seasonality Go 3.0.

With every new style choice adopted by Apple’s systems comes a collective sigh and opening of wallets from app developers as they try to not look outdated all over again.

Whilst this can be seen as a headache it is also a wonderful opportunity to refresh a brand and in turn breathe new life into an existing app.

With the update to iOS 8 Mike thought it would be a good time to take a look at updating and improving some assets surrounding the Seasonality Go app. We decided that the first place to start would be the app icon itself.

From a design point of view, the app icon is always a good place to start. It’s the gateway to the product itself and is the permanent representation of your product on the device of the user.

We wanted to include all the elements from the old icon but fit it to Apples new app icon development guidelines and give it that, clear, crisp feel. The first stage is always a more scattergun method to the process. Mike gave us the brief and it was up to us to develop as many initial ideas as we could think off. We varied the assets, composition, colours and styles, each time having the design stick as close to the iOS template as possible. It is through this initial stage of numerous preliminary ideas that conversations can be had and options refined.

From the too-ing and fro-ing of ideas, gradually, we arrived at the design we are proud to have now. A vibrant and bold design that captures the character and elements from the initial design whilst bringing the whole feel of the app into the iOS8 design era; even down to the Benjamin Franklin tribute key!

We hope you like the new design direction for Seasonality Go (and soon Seasonality Pro too) and share with us the exciting possibilities it brings in the future!

–

Chris Emery has 15 years experience designing and is the owner of cwench.com.

Cwench.com specialise in all things design, whether it’s digital, print or branding, we can help you and your business reflect the image you want.

Our clients find us flexible, friendly and professional so feel free to get in touch and tell us what you need!

[getseasonality]

Seasonality Go 3.0

I’ve been avoiding talking about this update on here for quite some time, but for the past few months I have been working on Seasonality Go 3.0.  Ordinarily, I would be fine talking about a new update, even a major update like a x.0, but this one was a little different.  It was the first Seasonality Go release to support the iPhone, and I wanted it to be a little bit of a surprise.

For years I’ve wanted a version of Seasonality to run on the iPhone.  I was never happy with any of the other weather apps I had used on that device, and I even had quite a few customers write in asking if I had plans to create a version of Seasonality for the iPhone.  

For some reason, I had it stuck in my head that an iPhone version of Seasonality Go had to be just as customizable as the iPad version, and I just couldn't figure out how to do that.  With Seasonality Go for the iPad offering a 4x4 grid, if I translated the size of each module directly to the iPhone I would be left with a 1x2 or a 1x3 grid.  I saw that as being kind of useless.

It wasn’t until after my mind started thinking outside the customizability box that it dawned on me that the content for an iPhone version of Seasonality Go already existed.  Each of the modules for the iPad version adapts to the size it’s being drawn at, from 1x1 to 4x4 and everywhere in between.  I just needed to design the interface to navigate between these modules on the iPhone.  

So I started brainstorming.  When I’m using an iPad, even if I am outside my home, I’m usually willing to spend more time working at a particular task.  Pulling out an iPad isn’t trivial, so if you are using an iPad chances are you are spending a reasonable amount of time with it before you plan to move on to something else.  Translating this idea to the usage of Seasonality Go on an iPad, it makes sense to let the user design their own weather screens and really give people a chance to explore the data in the way they best see fit.

For an iPhone though, I see things differently.  More than half the times I wake up my iPhone, I want to complete a task very quickly and then move on.  So I asked myself, how should a weather app be designed when the goal is checking the weather quickly, while still offering a significant amount of data if the user has more time to spare.  This is how I came up with the tab layout of Seasonality Go for the iPhone: Currently, Forecast, Map and Graphs.  

First, if the user has a few seconds, they probably just want to check out what it’s like right now and what it’s going to be like in the next few days.  For this, there is the Currently and Forecast tabs.  

If the user has a little bit more time (maybe because the current weather is affecting them in a significant way), then I imagine the next thing someone would like to see is a weather map.  Maybe they want to see radar data for a local storm, or quickly check the overall wind pattern using Particle Mode.

Finally, if something really special is happening, then Seasonality Go lets the user dive in to the graphs to really examine the data.  

It’s a fairly simple model, but I think it fits a lot of use cases.  I’ve already received a lot of great feedback, including some ways to improve upon it.  Software is always a work in progress, so there’s always more polishing that can be done to make Seasonality Go an even better tool for examining the weather.  

Check out Seasonality Go for your iPhone at the App Store.

[getseasonality]

Using Seasonality to Watch a Solar Eclipse

Tomorrow there will be a partial solar eclipse that will be visible from North America.  If you are in the Eastern or Central time zones, the eclipse will be happening right around sunset, making for a spectacular sight.  The problem for me, residing in central Michigan, is that an area of clouds will be moving in from the west right around the time when the eclipse is taking place.  I’m willing to drive a little bit to take advantage of a photo opportunity like this one, so I wanted to determine the location I am least likely to run into overcast skies during the sunset.

First I checked Seasonality Go.  Using the cloud cover graph, I am able to find out what the cloud cover is forecasted to be in a few different locations around my area.  The location I would most like to photograph the sunset from is to the west, on the coast of Lake Michigan.  Unfortunately, the cloud cover graph was showing about 75% cloud cover for tomorrow at sunset in that area.

Time to take a closer look.  I fired up Seasonality Pro to see what the various models are projecting for the total cloud cover.  I added total cloud cover layers for 4 different models: the GFS, NAM, HIRES (ARW Core), and the RAP.  

Instead of just looking at the cloud cover layers for each model individually, I wanted to composite them in a way that would give me more useful information on a whole.  Seasonality Pro allows you to customize the color gradient used for each layer.  So instead of using a clear to opaque white gradient for each model, I reduced the opacity for total cloud cover to about 25% for all 4 models.  This results in a display that shows opaque white only if all 4 models are forecasting 100% cloud cover.  Here's what I saw:

Looking at this composite, it’s clear (no pun intended) that the best place for me to view the eclipse would be in the eastern part of Michigan.  This was somewhat apparent from looking at the cloud cover graph in Seasonality Go.  However, in Seasonality Pro I can look at the specific areas where the least cloud cover is being modeled.  Since I would ideally like to see the sun set over a large body of water, I’m thinking that the west coast of the “thumb” of Michigan is probably my best option.

Even though Seasonality Pro isn’t quite done yet, it’s great to test out the app to see if it works well for real use cases during the development period.  In this situation, it worked perfectly.

[getseasonality]

Vertical Velocity

Some models like the HIRES and the RAP will output layers showing vertical velocity at different pressure layers.  Today I was exploring how to best display this, and decided on a blended layer with a gradient like one you might see for radial velocity radar images (green/- to red/+).  Here's a screenshot showing vertical velocity plotted along with the horizontal wind at 500mb.

It's interesting that you can see the effect terrain has on the vertical wind from this map.  The Rocky Mountains pass along the Idaho/Montana border, causing orographic lifting of the air that is noticeable all the way up at 500mb (around 15k-20k feet above sea level depending on atmospheric conditions).  

The patterns kind of look like spaghetti that is oriented perpendicular to the winds.  This makes sense, because as the wind passes up a mountain range, you have positive vertical velocity.  Likewise as the wind travels down along the back side of a mountain range, you'll see negative vertical velocity.

It's important to look at vertical velocity because it's a source of lift, which is one of the key ingredients of thunderstorm formation.  If you know where lift is taking place, you'll have a reasonable hint at where you might find clouds or thunderstorms spawning off as well.

At lower levels you could probably use this tool to see gravity waves and other vertically-oriented atmospheric phenomena.  I'll leave that for another day.

[getseasonality]

High Resolution Models

Some apps out there can show you a limited number of map layers from forecast models, but Seasonality Pro will give you full access to hundreds of variables in several different models.  In addition to supporting the common GFS and NAM models, Seasonality Pro also includes access to the HIRES, RAP, and Wave models (and I’m continually adding more to the list).

This allows you to really view some high resolution data that you can’t find anywhere else on a mobile device.  For instance, here’s a screenshot looking at simulated radar reflectivity from the HIRES model being run over the CONUS domain:

The HIRES runs in a very tight grid, with spacing of only around 4-5km, and this high resolution is available across the entire continental US (a total of over 2.3 million grid points).  You can zoom in to see more details in a particular region just by pinching on the map.  Since the map doesn’t use any bitmap images, the vector drawing using direct from the model results in a very clean display.

The drawback is memory consumption, and this is especially true on a mobile device like the iPad.  Each full-resolution dataset from the HIRES model takes about 45MB of memory.  This works fine on the most recent iPads, but slightly older models like the iPad 2 and 1st generation iPad mini can struggle a bit to keep up.  The plan is to have Seasonality Pro detect when a low-memory device is being used and compensate accordingly.  That’s next on my todo list. 

[getseasonality]

Plotting Multi-dimensional Data

For awhile now, Seasonality Pro has been successfully plotting single dimensional data.  I’m using single dimensional terminology to refer to data that just has a single value for every longitude/latitude grid point.  Single dimensional plotting is useful for a lot of situations, but sometimes you really need multi-dimensional data.  The most common case is when you want to look at something like wind speed.  In meteorology, wind speed is broken up into U (east-west) and V (north-south) components.  While you can certainly plot just the U component and get pretty close, since the majority of motion in the atmosphere is east-west oriented, it’s impossible to plot a completely accurate wind speed when you can only make use of one component.  

So it was time to add support for this.  I found it somewhat surprising that the code needed to add support for multi-dimensional data touched almost every layer of the application, from the user interface on top all the way down to the deepest layers of code used to download gridded data.  I added methods in the download code to help gather collections of data in the background, and only let the higher levels of the app know about it once all the data was finished downloading.  This will keep from having a bunch of duplicate code if other parts of the app needs to download multiple sets of data.

It took a few days to make all the changes and get things working smoothly, but the results are really nice.  It’s great to finally be able to plot a standard 250 mb chart and see the jet stream…