Hey, I'm Jordan Marr, and this is Farming in British Columbia, a podcast that frequently struggles to open feed and seed bags properly so that the string stitching into the top of the bag comes out smooth as butter. A process that we can all agree contains the greatest gulf between the satisfaction of getting it right and just watching and feeling that string zip on out and the rage-inducing frustration of getting it wrong. You know, just sitting there, frigging around with stitch after stitch as you stand there, hoping you're not the only one who gets it wrong. But fearing that actually dummy, you are literally the only one in British Columbia who can't easily open a sack of annual rye. Today on the podcast, a pretty fun topic that was suggested by listener Quentin Brunes. Quentin and his partner Daniella operate Hammerlin Holsteins, a dairy farm in Mera, in the North Okanagan. There are places right next to Wild Flight Farm, a market garden operated by Quentin's big brother, Herman, and his wife, Louise. Quentin contacted me to tell me that he and Herman disagree about the best source of weather forecasting. So I went out to Wild Flight Farm a few weeks back to interview the brothers about their competing forecast loyalties. And then I got on the horn with a couple of meteorologists to ask them why and how weather forecasts differ in the first place. I had fun with this one, and I learned a lot, and it is my sincere hope that both of those things happen to you when you listen to this. So we'll kick things off in Herman and Louise's living room, and then I'll introduce each meteorologist in turn. Talk to you in a bit. Alright, Herman Brunes and Quentin Brunes, thanks a lot for joining me on Farming in British Columbia. Yeah, it's interesting to be involved. Quentin, so you and Herman have farms essentially side by side out here in Mera in the North Okanagan, and you told me that you two have had disagreements over the superior source of weather forecasting for your area. So let's start there. Quentin, what is your favorite source of weather forecasting? Yeah, well, I definitely lean towards the weather network. I do like how they have their webpage set up. And there are certain bits of information that they include in their forecast that Environment Canada does not, and that's definitely pushed me in their direction. Ok, Herman, what do you prefer? Well, yeah, I'm the Environment Canada convert, or I prefer that. And basically, I think for similar reasons, they have most of the information I need. I will grant Quentin that weather network does have some, sometimes some information that Environment Canada does, but doesn't, but I love the clean, no advertising app. And it's just a lot faster and easier to get to. And the main thing too, I think I feel, if we want to get into it a little bit, is that I get the feeling that the weather network is more of a for-profit model and that they like to, or that maybe they've even structured their algorithm to make things more extreme. Like my experience is that if we're being called, heat waves being called for, they'll be like two, three, four degrees higher, or if it's a frost, they'll be like three, four degrees. And I feel like what they're trying to do is they're trying to generate headlines out of the weather. Weather drama. Yeah, weather drama. So that they can generate clicks and get people coming to their, because they've got to get people going with all the advertising that they've got on their website, they've got to get people coming to it. And that's sort of the feeling I sometimes get with them, is that they tend to be a little bit more extreme on things. And I feel like Environment Canada hasn't got those sort of competing interests. They just sort of present it in a scientific way. And that should be the cleanest and most unbiased weather around. So that's sort of one of my main reasons for choosing them. All right. And Kwanzaa, how do you feel about your brothers? Yeah, I think that's a fair criticism. I would concede that the weather network tends true to dramatize their forecast. A little more razzle dazzle. Yeah, yeah, for sure. But having said that, I don't mind that as long as the general trend is correct. I don't mind. Like, if there's going to be rain in the forecast, and I'm thinking about cutting some hay, I do want to be alarmed. And then, you know, when it comes down to it. So I really appreciate the fact that Weather Network does a 14-day forecast for me. That's really interesting information. Even if a lot of times are completely full of it. But it kind of signals a bit of a trend. And very often in the general trend, they're going to be right even if the details are out to lunch. So then I can, I will check Environment Canada. If they're both saying the same thing, then I'm feeling very confident that I'm good to go. But for me, the ultimate, the ultimate decider for me on preference for the Weather Network is the fact that they, when they give a percentage chance of rain, like 60% chance, they accompany that with the expected amount of precipitation. Whereas with Environment Canada, like 60% chance of showers. Well, man, it makes a big difference. Is that a millimeter or is that 20 millimeters? They don't give you a hint even. So like, I don't know why they don't have the balls to come out and make a prediction. Because it sounds like for you, then if you see that at EC, then you're heading over to Weather Network anyway to ask them how many millimeters. Right. Like, how much are we talking? That makes, like, I'll, I'll cut if it's a 60% chance of a millimeter big deal. But you know, I, 20 millimeters matters. Okay. So that, I wanted to establish how much the weather matters in your respective farm businesses. Because you do different farming. So I'll start with you, Quentin, like what, where does it, where does it really crucial for you? What parts of your farming? Oh, man, it is almost nonstop. It's almost nonstop, it's, it's kind of scary because there's, there's maybe two or three months of the year where checking the weather app isn't the first thing I do when I wake up. Yeah. Like it is really, it's like literally wake up, grab my phone. Okay. What are we looking at today? And that'll, you know, in some part determine on, you know, what I'm going to be doing that day or for the next couple of days. But in the, in the spring, you know, we've, we've got, I mean, you're spreading to do that you want to time with the, with the precipitation, summer, certainly hanging, which is kind of going on all summer long, fall, same deal where, you know, harvesting corn or wanting to spread manure again and have to be very careful about precipitation levels. And then, you know, it's a bit of a gap. And then it's winter time and like, okay, how much do we need to button up the barn so that things don't freeze in the caliber and our robot doesn't freeze up. So like I'm really, really keeping an eye on the, on the freezing temps as well. Okay. So I'm hearing high stakes a lot of the year for you. Sure. I'm really, so you're a dairy farmer, Herman, you grow vegetables, beautiful vegetables. I just had a tour of the farm. I'm really, I'm quite interested in your response because of my own experience growing vegetables and how, where I rated the importance of the weather forecast, where, how, how does it, how is it for you? Like, you know, give me, give me a sense of the difference between you check it because you're curious versus what the stakes are for your farming. Yeah. I would say I'm kind of in the same situation as Quentin that, yeah, one of the first things I do is look at the weather app because I want to know what's going on. Because, you know, if our team is arriving at the farm in the morning, I kind of have to have a general idea of what do we do in them with them that day. You know, we have lists of things that we do on particular days and sometimes we have to work around the weather. And so I think having the weather every day is, is, is, I would say maybe even more important because we're, we're out there every day, whereas Quentin's, Quentin's, he's just, you know, Calvin's tractor, yeah, we're out there physically, like reindeer and, you know, everything, whether it's raining or not where you have to be there. But, but see, but right there, that's why I wasn't sure where you were going to go with your answer because you have to be there anyway, whereas with Quentin, you know, it's going to sometimes dictate like, I am full stop, I'm not cutting the hay because that, you know, or whatever, you know, that's why I'll tell you. I'll just, I'll just confess because I actually feel a bit sheepish and guilty about this. Like you having heard what you said, you would be shocked as a veggie grower, how little I check the weather forecast, like, because I'm a sloppy farmer, let's, like, I'll just be honest, it's, it's, I, I should have been checking it way more and it was not part of my ritual and most of the time I did not know what was coming often to my detriment. Well, and at what, but, you know, I would qualify that in your case that you were working down south a bit more, especially in the summer, the weather is pretty predictable. Yeah. Yeah, we get a little bit more variable, we get more precipitation and that can really mess things up. If you want to be planting that day and it's going to be raining, you either got to get in there before before it says running because in our soils, as soon as it's, they start getting a little wet, it's game over for planting, just forget it, you got to get out of the field. You can harvest, but you can't plant or weed for that matter. And then certainly in, you know, in the shoulder season spring and fall, it's even more important with frosts and things like that. Are you opening greenhouses, closing greenhouses, you know, what's the high temperature going to be? How much cloud is there going to be those kinds of things are important, especially if you're going to be leaving the farm on weekends even, you know, we want to go and go for a hike or whatever. But oh, do we leave the doors of the greenhouse open or closed? And so you got to check, well, what's the high temperature going to be? How much cloud is there going to be in all that kind of stuff, right? So anyway, I find it's really quite important for us. I need to push back a little bit on the relative importance of checking the weather to each of our farms because I mean, certainly in the winter time, I don't think it's as critical to you. Like, the only reason you're checking the forecast in the winter time is to determine what the skiing conditions are going to be for you. So we also have to worry about greenhouses collapsing and markets. We do some farmers markets outside all winter long. So I'm looking at, oh, like, if we're going to Revelstoke, what are the road conditions going to be like? Is it, you know, if I'm sending one of my employees, maybe I should go along because it's going to be dicey or whatever. So, you know, if they're driving there and chances of avalanche and, yeah, I just sort of find that I'm looking at a fair bit in the winter, too, and I'm worried about our, you know, keeping on our greenhouse. You guys have convinced me that for any given topic I cover in the future on the podcast, I need to get siblings together to just like add an element of drama and tension. This is wonderful. Wonderful for me. Wonderful for clicks and listens. [beep] [beep] [beep] [beep] [beep] [beep] Good afternoon. Good afternoon, Armelle. How are you doing? Yeah, good. How's the farm doing? It's great. So right around the same time that I talked to Herman and Quentin, I interviewed Armelle Castellan with Environment Canada. Right. Armelle Castellan's and the title is Warning Preparedness Meteorologist with Environment Climate Change Canada. Armelle agreed to join me to do his best to answer some of the Brunds Brothers questions. Here's that conversation. Armelle Castellan, thank you so much for joining me on the Farming in British Columbia podcast. Oh, it's my pleasure. Thank you very much for hosting it. Okay, Armelle. So what I thought, like with this, we have this overarching question of kind of how and why do different agencies forecast sometimes differ. So I thought we'd start with just a pretty basic question about asking you to tell me and my listeners about the different tools and techniques that go into determining a weather forecast. So obviously, you'll be talking about what you know about what you know about Environment Canada. But maybe you could kind of give us a sense of how the forecast ultimately gets determined. There's many parts to a forecast. And of course, we intuitively maybe know that big computer programs are being used. But before you get to the actual model output, which has come a long, long way over the last, you know, several decades when computing has really kind of become such a powerful force, you also need to have a network of stations that are maintained and calibrated both on the ground. So whether it's at a lighthouse, on a boat that's moving around on land at airports, sometimes at elevation in the mountains and into 3D space. So we have every 12 hours of balloon will go up several places across BC, including our neighbors in Alaska, Washington State, and of course, into the into the prairies. And those balloons track and measure those very important elements like temperature, wind, humidity, all the way up the atmosphere until the balloon explodes way up at 10, 11 kilometers in altitude. And those pieces amount to an initial set of conditions that are extremely important to validate what what is the atmosphere even doing right now, as we can call it kind of the monitoring level where we have a baseline of actual observed conditions, whether it's the temperatures, as I said, the dew point, so which calculates the humidity in the air, how much is falling. So it could be snow and rain and sleet and so on, as well as the wind regime. And then you can imagine that that shapes a very three dimensional image that then gets ingested into the modeling and the models are there basically they're resolving a set of physical equations for the planet. And so models are often at the onset, global in nature. And so they can't be a hundred percent the size of, you know, an individual farm, let alone an individual house property. It's bigger than that. It's, you know, 10 by 10 or 15 by 15 kilometers in terms of the grid spacing. We then run that model to resolve all layers of the atmosphere in a supercomputer, which is based in Del Valle Quebec and push it out to 10 days. So that's the global model. We have more high resolution models that really go down to even one by one kilometer. It's called the high resolution deterministic forecast system and, but they only go out to 48 hours because, of course, when you reduce the size of the grid, it asks a lot of those supercomputers and then they, you can, you can't run it out to 10 days. So there's some, obviously some limitations as to what the models can do. And then this is where kind of the statement that makes a lot of sense to a meteorologist but may not be as common to the average person that all models are false. And the reason why that's an interesting statement is because we're just approximating what the atmosphere is doing, which is extremely chaotic of an environment. Obviously, there's room for improvement and things have been evolving. We've had essentially being able to add an entire day by decade of computing power over the last three, four decades. So, you know, back in the 80s, there's no way we were talking about the weather on day seven. A lot of the times, day seven is an impossible request out of the modeling capacity, let alone a human because humans do quite well forecasters, I mean, in predicting in the boat, the 48-hour window. Beyond that, often the computers are doing a better job and we, you know, give credence to what it can handle and what it can't. So I want to go back to your original summary real quick, like I just want to tell you what I took away and then I'm combining it with some things that I think I already knew, hopefully I'm accurate. But essentially, you really emphasize that there's a fundamental component of forecasting which is the actual measurement using real tools, not just algorithms, right? So those go up in the atmosphere and other ways of taking data. Then the data interacts with models and algorithms and then so where I'm going to riff on what I've thought I've understood before is that you're really taking data constantly and comparing it to what happened when we had similar data in the past. Like constantly taking the now and comparing it to what we know about what we could expect in the past, that's kind of how the models work and then we're constantly refining those models. Is that some version of kind of sort of correct that kind of thing? Yeah, no, I like that way of explaining it, Jordan, I think you are reanalyzing the atmosphere continually. Some stations are by minute because you're getting a lot of precipitation and it would be good to know if you're getting 100 millimeters an hour for five or 10 or 15 minutes and other times the information, as I mentioned, with the weather balloons is every 12 hours. But absolutely the sensing of the atmosphere informs the modeling and reinitializes it on a continual basis. You're kind of recalibrating, calibrating is a fun word because it just means that you're course correcting what you know of the atmosphere to what the model can then provide and then of course you'll never get out of the interpretation of that model output so that those computer models do a bang up job but they are not perfect and in fact what we'll do is compare them to one another. So the Americans have a lot of models that we have access to, the Europeans, the Japanese, the French, the UK, all have big meteorological agencies that run their models certainly at the global scale all the time and you know we share our models with those agencies and vice versa and we can compare them and make really informed decisions on whether or not a model or two and it could be the Canadian is kind of out to lunch what we call an outlier and therefore needs to be respected but not necessarily weighted any stronger, in fact probably weighted less strong when it is an outlier and then we go forward from there. It's so interesting because it's like if I take the super computer out of it I might have assumed if we were just talking about humans forecasting I might have assumed that just on a point of pride you know meteorologists are going to want to constantly be reviewing forecasts to determine how accurate they are or were in order to become better but you're just saying it's like it's a fundamental part of the system is the constant analysis of how accurate the forecast was as these models are being continually refined. Yeah I would go further and to add a social science element to the metaphor of the models being re-initialized all the time and getting better for the next forecast period but also apply that to a forecaster I mean there are studies that show that forecasters forecast better for where they are located because they are continually being fed the feedback of what the atmosphere is doing in their own senses with their eyes their temperature you know their skin in a location where there's so many more nuances at the meso and micro scale then you know we can forecast specifically so you know we human brains are capable of pattern recognition and that's a big part of what we're talking about here with humans are very much a part of the equation to put the pun in there and I think they'll stay that way I don't I'm not afraid for me or my colleagues jobs it's certainly not into the medium term here. Earlier at the very start of your summary of the tools you talked about fancy measuring tools like balloons and other things if we doubled the amount of measuring tools in other in other words doubled the density of of the location of these tools would would that have quite a corresponding increase to accuracy is that a good way to think about it or what are the limitations of thinking about it that way. Absolutely I mean Canada is five and a half time zones it's just absolutely enormous we have different challenges than we do say in Switzerland forecasting the weather you know kind of a postage stamp in comparison we don't have a density that is very enviable of more southern countries where access and capacity to implement you know a stations right across the country. So if we had a lot more we do better and a great example of that is the Coco Rass Network C-O-C-O-R-A-H-S.org is a kind of a citizen science driven participatory collaborative community and so people who participate are able to measure every 24 hours and add comments into a database that's then used both academically and also in near real time because at seven or eight a.m. when all of that information is ingested and into the into their computer system you know forecasters use that and will look at and verify precipitation estimates were they on board and they compare well to the official airport stations for instance and you know you you can always use extra data especially when it's relatively coordinated. All right so let's go back I want to take us back to this kind of main thread or main question here so I think you've given us some good context to start out with Armelle like how or why might two agencies that offer forecasts for a given region how or why might they differ what what goes into having them predict different weather in a given region. You know it all depends on who's producing it and it might not even be a who it might just be a what and so if it's the stock weather application that you find on any given smartphone chances are it's just raw model output as we call it so it's not tampered with by a human because it's just absolutely just kind of automated algorithm that will scrape for you know a spot that you're interested in but those that output will routinely be easy to ridicule ridiculous because you know I will wake up here in my hometown of Victoria and you know compare what is being offered to me as a meteorologist I find that quite interesting and I'll often in the middle of winter be told okay well it's a minus 17 and you're going to get to a high of minus five and in Victoria that is not only unlikely but almost impossible and it it can be really an interesting way to understand why raw model output is not always very helpful you can imagine that the models will have a trickier time dealing with all the micro climates of a place like British Columbia because we have you know many mountain chains and obviously a big area and an ocean on one side and sometimes the influence of continental Canada coming in from the north or from the from the east and and so that environment makes it extremely hard for raw models to make any sense of of their output so beyond that you know we have at Environment Canada sometimes three overnight sometimes four or even five forecasters who are only responsible for the BC and Yukon landmass which is obviously very different from other agencies which may have fewer or international eyes on it as opposed to local and and the big big difference of course is that we at Environment Canada have a mandate to issue alerts for public safety 24 hours a day seven days a week stats holidays everything and and and nobody else does and so we are issuing you know could it be wind related warnings heat related warnings smoke related alerts rainfall snowfall colds arctic outflows you name it those are the things that are going to be a very big headache for people who don't know they're coming and give people that type of very deliberate heads up on whether that is going to absolutely be related to public safety so our mail I told the two brothers in question that I was going to be talking to you and I asked them if they had any questions they submitted three that I have here they didn't submit them with their voices they use text so I used a voice emulator do you mind if I play them for you real quick oh go ahead yeah awesome okay so the first one is from Quentin although this isn't Quentin's voice as will be pretty obvious so I'm gonna I mean you're gonna listen to this question now why does an environment Canada include the actual amount of precipitation expected in their forecast it would be really helpful to know whether there's a 30% chance of two millimeters or a 30% chance of 20 millimeters so does that ring a bell to you I went and confirmed I found that at least the hourly forecast wasn't including how many millimeters when it gave a per per possibility of precipitation so it's a great question I appreciated a lot yeah I mean first of all chances of of precipitation is fairly misunderstood and we can get back to that after but the the fact of forecasting precept whether it's rain or freezing rain or snow hail etc is extremely hard so it's much harder than predicting temperatures you know seven days out even tomorrow's convective elements in the interior of BC are going to be much harder to get quite accurate and you could also say that you know in the middle of winter on the coast with the big atmospheric river event also has a lot of terrain driven features that are going to give you a big range in terms of what you might expect so for any one location when we start getting above 10 millimeters of rain which is you know a sizable amount like it's definitely gonna be a wet day out there if you're trying to ride your bike then you or if you're farming you'll definitely see the impacts of 10 millimeters of rain or more on the ground whereas maybe you know a trace or a couple millimeters may not have a big impact if it's nice and sunny out so when the quantities become part of the equation that part of the forecast then we what we do include them in kind of 12 hour increments so we'll talk about today being about 12 hours tonight being another 12 hours and so those part periods will include the amount of precipitation however we usually don't include precipitation beyond 48 hours because there's just too much chaos to be realistic about it and it's better not to give the public or the users of the forecast falsely leading information really because you might that might change quite a bit it might go from a big strong bullseye as we're seeing right now for Fort Nelson you know next Wednesday but you know I'm not going to advertise 200 millimeters that one or two models are indicating because by the time we get to you know Monday or Tuesday next week that forecast will have dramatically changed to likely less in this case and we wouldn't want to raise alarms by you know starting to talk about the wildly chaotic nature of precipitation forecast you know beyond about 48 hours just back to my one of my first principles in this conversation was about where's the room to expand on and evolve into with the public similar to what a lot of the European countries have evolved on and and whether ready nation down in in the United States is to really help provide a more probabilistic language and not as much deterministic language so maybe the better approach to precipitation forecasting back to the question of one of the brothers there is to maybe give a better a range and to give a sense of where things are more fuzzy and where they're more concrete and that I think there's obviously room for expansion as we go forward in time yeah I think like quite I happen Quentin's a dairy farmer and so he's no doubt it becomes most crucial for him when he's cutting hay and needs it to get dry and I think that's where his question comes from in terms of like boy would it help to know when I see that percent likelihood of precipitation whether we're talking about minute amount or a serious rain I think that's where he's coming from yeah yeah do you want to take this chance to debunk some or like talk about misunderstandings of the phrase you know POP probability of precipitation yeah so if I'm in a given region that a forecast is reporting on and it could be a large region like the north Thompson right yeah and there's a 30% chance of rain tell me tell me what that means for me in that region yeah so again yeah it's and I've had this conversation with media over the years and it's it's it's amazing because it's something that we put in our forecasts all the time and yet it's often misunderstood so in the case of you're in the north Thompson somewhere whether you're in you know clear water or something like that and you see the forecast giving you 30% chance of rain you know say it's for the afternoon you're essentially being told that the atmosphere has a pretty good chance but it's not a guarantee to produce some rain or produce some snow for that period of time and we are estimating that 30 30% of those chances are that it will bring those point one or point two millimeters or centimeters of the respective precipitation so there's actually in that case a much larger percentage chance that you will not register that amount and that's that's kind of the difference between a forecast of 30 pop or up in 70 pop where you're kind of saying that the atmosphere when it behaves this way when we forecast it for the atmosphere of this nature in the past we are relatively certain in that case 70% chance is a much higher chance than 30 that rain or snow is going to fall but I get that point part but what I don't get is if for the 30% example are you simply saying there's a 30% chance that somewhere in the region in that reporting region it will rain not not the whole region not the whole region absolutely I got and that's also yeah absolutely it's not region specific in in a in an area forecast or a forecast region it's different in the United States where they will they have a gridded forecast so you can click on any little tiny grid all over the the continental United States and it will pump out a pop for that particular grid spacing and so that's a slightly different version of the you know that's a precipitation that we use in Canada so if they're a little bit or they're definitely similar but they're not exactly the same and so they shouldn't be compared because they're not doing forecast regions for their for their outputs right can I play you another question from one of the brothers yeah okay this is this is from Herman although this as you will definitely no understand is not Herman's voice okay the EC weather app provides a weather forecast for Mara but as far as I know the closest stations where weather is recorded are in salmon arm and Vernon they are probably using models to automatically generate the forecast but sometimes the weather for Mara is quite different than for salmon arm or Vernon maybe they could go into how the weather model works and whether the forecast for towns with actual recording stations are more accurate than for locations that are 30 to 50 kilometers away since the valley gets quite a bit narrower toward Mara and north there are times will get different weather like more cloud and precipitation than Vernon or salmon arm there are times where I'll compare Vernon salmon arm and Mara and make my own estimate for what the temp and precip could be in Mara so that's a pretty general one but any comments yeah I think I mean that's the reality we went we talked about it early in our conversation is the the the density of the network is is only what it is and add all the specificities of every last you know unincorporated small parts of a forecast region in places that have their own version of you know the thousand micro climates of BC will know instinctively just like this farmer friend of yours knows that they need to adjust the forecast for their area if you know they routinely go from Mara into salmon arm and Vernon and see that you know fog develops along the highway near enderby more than salmon arm or that you know the the proximity to the lake adds an element of moisture that is going to be stronger in one spot than the other then our forecast is just impossible to give those details even though you know we can you can look at the the the radar signature because we have radar in that area you can look at the satellite image and kind of pick pick out the the valleys that get more more cloud than others vice versa so there's yeah there's an end degree that you can quite going to quite strong details about one forecast versus the other and I mean that's part of why the private industry exists that there's some need for spot forecasting that various industry requires that is beyond what the public forecast is is able to do and sometimes those agencies whether it's kind of academically driven or for a baseline whether it's kind of environmental consulting they might have to actually put down some instruments as a baseline for months sometimes even years in order to be able to provide a forecast or to provide you know information for the user in those specific settings are those there anything else you'd like farmers to know about your work or the work at Environment Canada before we say goodbye yes I would just like them to understand that our forecasts are all manually adjusted by trained professional meteorologists and none of your phone apps are going to do that and so this is where we we have a very big difference we know and we take that very seriously and we put a lot of scientific energy into our forecasting system and that's that's the last thing I would like them to know our Mel Castellan thank you so much for being extremely generous with your time for an aspect of your job that is not the direct point of having someone like you who can get on the phone with journalists you normally you're I understand you're often talking about hazardous weather and impacts of weather not just answering dumb questions from a podcaster so I really appreciate it well I don't I don't believe for a second these are dumb questions I think it's a hundred percent part of the outreach that is necessary to really kind of keep the the population educated and uh and along for the ride because there will be changes over the coming decades our Mel thanks a lot thanks very much Jordan take care so that's what our Mel at Environment Canada had to say you may have noticed that he didn't offer any disparagement of the weather network both our Mel and my next guest Doug Gillum from the weather network stressed that they had no interest in slagging on each other in their interviews so I refrain from encouraging the kind of slander you've all come to expect on farming in British Columbia with that said here's my interview with Doug Gillum a meteorologist who manages the forecast center for the weather network at its headquarters in Oakville Ontario Doug Gillum thanks so much for joining me on the farming in British Columbia podcast hey my pleasure great to join you Doug our why might two different weather forecasting agencies present different forecasts for a given region how do we sometimes end up with with with forecast that vary for say the North Okanagan region in British Columbia well it depends on the time frame we're looking at but the simple answer is I mean that made the reality is meteorology is an inexact science at this point and really will will always be that way there are so many variables that impact the the ultimate forecast it's not just plug the numbers into an equation and outcomes forecast that we know will verify perfectly so you know there's different approaches to forecasting I mean many other outlets are going to use strictly a computer derived forecast there are computer programs that you know they are they will put out the forecast without any human involvement whereas you know one of the things that we do and also Environment Canada does as well as there's a human behind the forecast outlets that just use a computer derived forecast won't be able to make those judgment calls but they are judgment calls and you know you can you can get two different opinions by looking at the the same data when you're when you're an agency as large as the weather network how does it work like you can't I would I'm guessing you can't possibly have people spread let's just talk about Canada like throughout the country in all these different regions so how like with the weather network how how regional does your do do boots on the ground get if at all are you is it centralized in in one part of Canada and then it's just the data that you're taking from across Canada or or how how does that work our office is in Oakville which is just outside of Toronto and we produced forecast for all of Canada out of that office there's dozens of models that produce forecast data and we look at that data and we bring that data into our forecast system and then we can do manual adjustments and you know you know we'll forecast for all of Canada that way but is you know we make decisions at varying levels like we may decide you know for this current setup this particular model has the best handle on southern BC weather and so we'd go ahead and put in that model but then you go and you know we can go all the way down to 10 kilometer pixels and you know so it may be there's a pixel for Kelowna pixel for Vernon pixel for you know many different communities in the area and you can go and make adjustments and you know we're looking at sat you know one of the things that might you know you really want to adjust for it you know one or two degrees in temperature doesn't usually make a break for most individuals but we can look at satellite so that's looking down on a region and we can see that well the model said it would be clear by now but we've got clouds you're just stuck in the Okanagan valley and so we can go in and just expand the cloud cover or increase the duration of the cloud cover through all the regions impacted by that so satellite allows us to look down and see where it's clear versus where it's cloudy and radar allows us to see where there's precipitation occurring versus where it's not now in mountainous terrain the radar I mean the mountains do create a problem because you could be experiencing drizzle in the valley and the radar is just not seeing that so that's where you have to try to get some ground true somewhere to recognize and maybe the forecast has gone off the rails how about back further up the process are the major agencies effectively using the same actual data how does that work like to what to what sense is gathered data proprietary versus you know gathered for everyone in the game everybody's depending on an environment Canada for for those conditions they collect the data they send out the data and that's available equally to us as it is to to really anybody in the world now in terms of the the data that you get from the balloons that are are sent up to give us conditions in the you know up where we don't live but where the weather happens up in the atmosphere that data is collected you know those balloons are launched at the same time twice a day around the world and that data is collected and disseminated by a global the world meteorological organization and that data is again available to you know it gets ingested into all the models that are run now some of the model data you have to pay for and you know we we we pay for that data but you know that data if you're willing to pay for it is available you know to agencies around the world I want to ask you a question that came directly from one of the the farmer the farmer brothers that generated the idea for this story because one of them really favors environment Canada's forecast and one of them really prefers the weather network forecast and one thing that the brother the brother who prefers the weather network one thing he appreciates is that generally speaking when the weather network is is is forecasting precipitation they're more likely than environment Canada to actually estimate an amount of precipitation because he you know for example he produces hey and you know if there's a if there's a 60% possibility of precipitation okay is it going to be for two millimeters or 12 millimeters because he can probably handle two millimeters if he's going to decide to to cut hay but but 12 millimeters is a is a different story altogether why is that do you do why why do you take that tack at the weather network you know when we produce a forecast we're always I mean the amount of precipitation's a critical component of that so that's something that we're always generating as part of the forecast process and the question is then it comes down to your platform and your app and your website what do you what ends up being public and what ends up not being public and you know that's something that we put out there because of the you know because there is value in that there is there are people who want that information you know we're a media outlet and and so you know our our our mission is to provide the information that's beneficial and valuable to the public so you know we can put a little bit more of our time and resources into developing a platform that will be beneficial to the you know provide the public what they want of course as a meteorologist I recognize that especially in this critical time of year you know during the growing season where so much of our precipitation is scattered showers and thunderstorms you know you know the meteorologist sometimes we kind of cringe at the fact that a number is going public because we know that with you know thunderstorms are are so hit and miss and the reality is you know that two or 12 millimeters may represent sort of an average of what we would see in that area and yet you know on a day where there's popcorn showers and thunderstorms the expression when it rains at pours is very applicable and those who are impacted by the thunderstorm may get 25 millimeters whereas two kilometers away you get zero and that becomes very difficult to represent because you know the meter as a meteorologist I know that that number could verify wells and average for that 10 by 10 grid and yet the individual farmer experience is unlikely to get exactly that amount you know we've all seen times where even across a large you know or across a small community the amount of rain from one end of the community to the other can can be significantly different I have gone down the street and it's pouring and it wasn't raining at all at my house and so or vice versa so that type of detail becomes well reality it's kind of impossible to you know with much lead time to communicate on on an app or a website but then okay I want to link that to something you said almost at the outside of this conversation which has to do with um expectations around accuracy you know one of your first comments was that it's it's it's it's it's really hard to be uh super accurate it's just like you know it almost I don't know I almost think of like what what what agencies accomplish as kind of magic and I'm just wondering do you think the public has developed over the decades just unrealistic expectations around what should be expected for accuracy given all of the variables involved uh yes I mean I love the fact that the public expects accuracy because you know I mean because we we strive for accuracy and I think the fact that the public's expectations are so high now speaks to the progress of the science uh what we you would have considered to be accurate 25 years ago is not anymore I mean people want to know exactly when it's going to rain in their backyard and I think that that expectation wasn't there 25 years ago but the progress that has been made is you know we always want to be better and we always want what's being provided to us to be better and I think as forecasts have become more accurate and and as we deliver more details I mean you couldn't I mean we 25 years ago you didn't have an app or a website and so you I mean that to you just you just had a forecast on the radio that said scattered showers and thunderstorms for the next couple days um you didn't there was nothing trying to tell you exactly when it would occur and how much and so as forecasts have gotten better and as more details have been provided with that has come an expectation which we strive to reach but the reality is I can't tell you exactly how much rain will fall on in your backyard on a particular day when it's hit and miss showers and thunderstorms I wish we could but that really is beyond what the science probably ever will allow but we will always strive to be better Doug can you talk about the difference in the challenge of forecasting when you're talking about a fairly mountainous part of the country like British Columbia and specifically say the North Okanagan and as that blends into the foothills of the Rockies versus out on the prairies say in in in Manitoba or Saskatchewan does does the does the does the does the challenge become much greater in terms of accuracy when we're talking about BC absolutely you know just with the range and terrain over short distances one the models aren't going to handle that well to the when you have I mean you can have such sharp contrasting conditions over short distances in so many local effects that you know when you're you're looking at Edmonton you know that the terrain is pretty consistent so you know you don't have you know a wide range in conditions over short distances due to mountains and valleys on the flip side you know scattered thunderstorms still occur across the prairies and it's you know whereas in bees you know well it's very difficult to nail the temperature when the there's so much change in elevation over short distances on the flip side the terrain can cause consistent tendencies that become more forecastable as you get to learn the area and this is where you just you have to beat the models and take the time to learn an area and this is where reality is a local farmer who is observing the weather can beat the forecast in a short term because they know that the thunderstorms typically miss their field or typically hit hit their property because of local effects but so the terrain can cause some consistency in the randomness whereas you get out onto the prairies they still get scattered thunderstorms and you're just left going well I don't like it's watching a a pot that's about to boil where's the first bubble going to come up I don't know it's totally random you know the bubble is going to come up and then there's going to be lots of them but without terrain to kind of give a trigger you don't know where the scattered storms are going to be whereas in BC for someone who's willing you know who's able to really take the time to learn how the terrain is impacting the weather you can pick up on local effects that are not forecast by the models that are still predictable because of their consistency and and they make sense as you know the terrain right but that's a pattern recognition yeah and I guess there's always a risk of confirmation bias but if you're careful then something like you know I've observed that I live in a rain shadow because of those mountain that those these mountains over here can really can really actually be a reliable tool in in a farmer assessing their own risk in their own forecasting absolutely farmers have to pay attention to their weather their livelihood depends on it and someone who is a keen observer and you know maybe using what is forecast but can recognize under certain setups I usually get hit in this type of setup or I usually get missed on a keen local observer can can really use the forecast and then do even better to know how it applies to their property or to their farm as they you know as they learn pattern recognition is there anything else you hoped to get across that we didn't cover we've talked a lot about I think just kind of like is it kind of you know today's forecast tomorrow's forecast you know I get asked a lot about so why is our summer forecast different than another agency and the answer is slightly different there the environment can't it's very open about the fact that there's summer for their seasonal forecast is just from a model whereas I we look at models but we use a different approach yes we look at years in the past that had similar global patterns they're called analog years I mean that's just more accurate than the models the key is did you pick the right years but we can diagnose years where this year is going to be characterized by extreme heat or extreme drought so that's a different scale of forecasting and a different approach to forecasting but one there's increasing interest in that you know if a farmer knows that this is a year where drought's a bigger concern or extreme heat's a bigger concern or you know we're going to be dealing with a lot of rain this year they can actually make some decisions going into the season based on that and the hope would be if you're making those decisions there'd be some accuracy and results from that right and it sounds like you take you take pride in in the weather networks job that it does on those long range forecasts absolutely I mean that's kind of I mean it's I spent a lot of time on that I take it very seriously I fully understand the limitations of it but generally I take pride in being able to diagnose the key drivers of a pattern for a season and trying to give information that could be useful especially to you know to farmers because they're making this important decisions you're not going to change your weak invocation plans around the seasonal forecast but you might change you know might impact your approach to the growing season just on my own farm it can make you know if I if I care to pay attention to that it can help me decide whether to plant a short season maturing corn or a long season maturing corn yeah you know that's that's important yes yes exactly and that's the type of thing we would hope to be able to help with because we you know we appreciate our farmers and if we can help them you know that that's great that's a win-win Doug Gillan thanks so much for taking time to join me on farming in British Columbia my pleasure take care well some fun news is that I was able to secure an interview with a meteorologist from environment Canada so I this has been really interesting listening to you because I have this impulse to actually start revealing to you right now some of the things because you've actually brought up stuff that we talked about in the interview I shared the questions that you submitted to me with the meteorologist but I think I'm going to exercise some restraint and not not talk about that right now because you guys can listen with everyone else as to what what what our mail from Environment Canada had to say I will will fit will wrap it up here I'll just ask you um you've touched on them already probably a little bit but like any any any weather forecasting gripes you want to get off your chest right now we've heard one from Quentin about the lack of a POP number an actual like amount of preset from Environment Canada anything else yeah I think I share that frustration sometimes too I kind of wonder how they figure out the percentage of precipitation in the first place and I mean that is sort of a an nebulous thing like today it was supposed to be 60% chance to shower as well okay we got like three drops does that qualify for 60% chance here is that like 100% like I don't know so uh whereas environment uh um weather net was was seeing 40% chance and less than one millimeter for each of the time periods um so it was kind of like six and one half doesn't other they were kind of very close today well this is great you're given props to to weather network and and I got to give props to Environment Canada because I I checked Environment Canada forecast this morning and I see this thunderstorm warning and I was like what the hell it was it was just on the weather network you didn't say anything I go back to the weather weather network there's nothing and and for Environment Canada severe thunderstorm watch what the hell this is like how can you not have it and sure enough we had this wicked windstorm so um I really think we need to end it there that you too just acknowledge the other's agency it does a good job uh your listeners can't see that that was that was Quentin given his bigger brother a big kiss on the cheek this is really nice guy guys good vibes uh Quentin and Herman thanks so much uh well Quentin thanks for the idea I'm looking forward to sharing this and I really can't wait to hear your um feedback on the interview with Armelle thanks a lot guys thank you thank you okay that's it for now other than saying that Quentin's original request for this episode was that I asked other farmers their preferred source of weather forecasting which I went ahead and completely ignored so if you're grateful to Quentin for inspiring this episode maybe you'll consider sending me a voice memo to tell me what forecasting you trust instructions for sending me a memo are in the show notes of every episode but here's my number 250-767-6636 you can text or whatsapp or signal me or email a voice memo to podcast@farminginbc.ca and hey I get it you'd like to send me something but that would take you two or three minutes of effort I can relate because it took me six to eight hours to produce this episode and many of those were spent editing in my garage and it's real lonely in here so please farmer won't you tell me who tells you when it's gonna rain that's it for now talk to you soon everyone and remember we have more in common than all our differences would suggest [Music]
