WEBVTT 00:00:00.000 --> 00:00:02.960 Okay, imagine this. Picture it playing out, maybe 00:00:02.960 --> 00:00:05.919 even right now. You've got this powerful hurricane, 00:00:06.160 --> 00:00:10.080 just this massive swirling thing, and it's heading 00:00:10.080 --> 00:00:12.660 straight for the coast, densely populated coast. 00:00:12.800 --> 00:00:14.779 Millions of people. Millions, and they're all 00:00:14.779 --> 00:00:17.719 hanging on, you know, bracing themselves, relying 00:00:17.719 --> 00:00:20.980 completely on those forecasts. Absolute precision 00:00:20.980 --> 00:00:24.429 needed for, well... life -saving decisions, evacuate, 00:00:24.989 --> 00:00:28.269 shelter, and then boom, just like that, no warning, 00:00:28.469 --> 00:00:31.230 one of our most critical eyes in the sky, this 00:00:31.230 --> 00:00:34.689 vital window into the storm's heart. It just 00:00:34.689 --> 00:00:37.869 goes dark. Just switched off. Switched off. Welcome, 00:00:37.909 --> 00:00:39.929 everyone, to Meteorology Matters. We're your 00:00:39.929 --> 00:00:42.590 guides today, and our mission, like always, is 00:00:42.590 --> 00:00:44.689 to take this, well, complex, sometimes really 00:00:44.689 --> 00:00:47.189 daunting information from all sorts of places 00:00:47.189 --> 00:00:49.590 and really distill it down, make it clear, give 00:00:49.590 --> 00:00:52.130 you actionable insights. Yeah. The goal is to 00:00:52.130 --> 00:00:54.829 help you become genuinely well -informed, not 00:00:54.829 --> 00:00:56.649 just the headlines, you know, but the actual 00:00:56.649 --> 00:00:59.670 science -shaping things. Exactly. And today, 00:00:59.750 --> 00:01:03.429 we're digging into a really sudden, very impactful... 00:01:03.369 --> 00:01:05.849 honestly, kind of mysterious decision came down 00:01:05.849 --> 00:01:07.969 just this week from the United States Department 00:01:07.969 --> 00:01:10.349 of Defense. Unexpected, is putting it mildly. 00:01:10.590 --> 00:01:13.909 They've decided to just terminate the processing 00:01:13.909 --> 00:01:17.709 and delivery of absolutely vital data, data streaming 00:01:17.709 --> 00:01:19.890 from a specific set of weather satellites. We're 00:01:19.890 --> 00:01:22.810 talking about the Special Sensor Microwave Imager 00:01:22.810 --> 00:01:25.750 Sounder, or everyone calls it SSMIS. Yes, SSMIS, 00:01:25.810 --> 00:01:28.069 yep. That's the one. And let's be crystal clear 00:01:28.069 --> 00:01:30.609 here. This isn't some, you know, minor technical 00:01:30.609 --> 00:01:32.670 tweak only specialists care about. No, no, this 00:01:32.670 --> 00:01:36.189 is profound. It's like a a seismic shift in how 00:01:36.189 --> 00:01:38.209 we observe these storms. Absolutely. It's a move 00:01:38.209 --> 00:01:41.290 that's going to severely impede significantly 00:01:41.290 --> 00:01:43.969 degrade hurricane forecasts. And not just for 00:01:43.969 --> 00:01:45.549 this year, the twenty twenty five season that's 00:01:45.549 --> 00:01:48.620 starting now, but. way beyond future seasons, 00:01:48.760 --> 00:01:51.819 too. We're talking direct impacts, tangible tens 00:01:51.819 --> 00:01:53.959 of millions of people living along those coastlines, 00:01:54.079 --> 00:01:56.439 their safety, their economic well -being. It 00:01:56.439 --> 00:01:58.959 relies on accurate predictions. And what's really 00:01:58.959 --> 00:02:01.359 baffling, truly perplexing is the lack of any 00:02:01.359 --> 00:02:04.519 immediate, clear reason. Why now? Why so abrupt? 00:02:04.780 --> 00:02:06.819 It seems to be tangled up with Department of 00:02:06.819 --> 00:02:09.620 Defense security concerns. But the really alarming 00:02:09.620 --> 00:02:12.620 part, even top officials at the National Hurricane 00:02:12.620 --> 00:02:15.759 Center were completely blindsided. No heads up, 00:02:15.960 --> 00:02:18.280 nothing. caught totally off guard. Yeah, it's 00:02:18.280 --> 00:02:21.460 just wild. It's unprecedented, really. And the 00:02:21.460 --> 00:02:23.500 ripples are already spreading through the whole 00:02:23.500 --> 00:02:27.360 weather and climate community. So over the next 00:02:27.360 --> 00:02:29.159 little while here on Meteorology Matters, our 00:02:29.159 --> 00:02:33.060 job is to dive deep, figure out exactly why this 00:02:33.060 --> 00:02:36.319 data from SSMIS, why it's so uniquely vital. 00:02:36.439 --> 00:02:38.120 Right. What are the concrete impacts going to 00:02:38.120 --> 00:02:41.219 be? It's loss on understanding, tracking, preparing 00:02:41.219 --> 00:02:43.800 for these massive weather events, hurricanes, 00:02:44.099 --> 00:02:46.360 climate patterns. And critically, What's the 00:02:46.360 --> 00:02:48.819 path forward? Because the global meteorological 00:02:48.819 --> 00:02:51.000 community is scrambling now, almost frantically, 00:02:51.479 --> 00:02:54.039 trying to adapt to this sudden hole in the data. 00:02:54.400 --> 00:02:57.360 So let's unpack it. Absolutely. Let's peel it 00:02:57.360 --> 00:02:59.360 back. Because to really get the gravity here, 00:02:59.500 --> 00:03:03.080 you first need to appreciate the, well, the absolutely 00:03:03.080 --> 00:03:05.780 essential role microwave satellite data plays 00:03:05.780 --> 00:03:08.199 in forecasting hurricanes. It's intricate stuff. 00:03:08.780 --> 00:03:11.319 And hurricanes themselves, they present this 00:03:11.319 --> 00:03:15.120 fundamental challenge for observation. It's plagued 00:03:15.120 --> 00:03:18.340 meteorologists for well forever decades. Yeah, 00:03:18.780 --> 00:03:22.039 these huge storms They mostly form and get stronger 00:03:22.039 --> 00:03:24.699 over these massive stretches of open water. Think 00:03:24.699 --> 00:03:28.270 about the scale hundreds thousands of miles from 00:03:28.270 --> 00:03:30.530 land. Where you just can't put ground measurements. 00:03:30.569 --> 00:03:32.870 It's not feasible. Right. No weather stations 00:03:32.870 --> 00:03:35.270 out in the middle of the Pacific or dotted across 00:03:35.270 --> 00:03:37.969 the deep Atlantic. We just don't have them. Exactly. 00:03:38.050 --> 00:03:40.189 That's the core dilemma. And yeah, we have the 00:03:40.189 --> 00:03:42.229 Hurricane Hunter planes, those amazing brave 00:03:42.229 --> 00:03:44.610 crews flying right into the eye. Incredible data 00:03:44.610 --> 00:03:47.789 from them. Invaluable, but they have real practical 00:03:47.789 --> 00:03:49.969 limits. These planes, they're only available 00:03:49.969 --> 00:03:52.750 for about one out of every three Atlantic hurricane 00:03:52.750 --> 00:03:56.379 forecasts. One in three. Yeah, think about that. 00:03:56.620 --> 00:03:58.780 Two -thirds of the time, we don't get that direct 00:03:58.780 --> 00:04:01.960 granular data from inside. And for Pacific storms, 00:04:02.639 --> 00:04:05.259 almost never. Maybe a handful of the really extreme 00:04:05.259 --> 00:04:08.340 cases. So most of the time, for these huge ocean 00:04:08.340 --> 00:04:11.099 storms without those planes, we're kind of flying 00:04:11.099 --> 00:04:14.280 blind, making critical calls with way less information. 00:04:14.740 --> 00:04:16.939 That's absolutely, yeah. It leaves this huge 00:04:16.939 --> 00:04:20.600 observational gap, a void that desperately needs 00:04:20.600 --> 00:04:23.420 filling for accurate forecasts. And that's where 00:04:23.420 --> 00:04:25.699 the satellites come in. But here's where it gets 00:04:25.699 --> 00:04:28.720 really interesting, where the power of this specific 00:04:28.720 --> 00:04:31.579 data, the stuff we're losing, really shines. 00:04:31.680 --> 00:04:33.899 There's a key difference, right, between the 00:04:33.899 --> 00:04:35.759 satellites most people are familiar with, the 00:04:35.759 --> 00:04:38.819 ones giving those beautiful swirling cloud pictures. 00:04:38.980 --> 00:04:41.259 Yeah, the geosynchronous ones you see on TV. 00:04:41.519 --> 00:04:44.920 Exactly. Those versus the more specialized polar 00:04:44.920 --> 00:04:47.120 orbiting satellites carrying these microwave 00:04:47.120 --> 00:04:49.779 sensors. OK, so what's the difference? Those 00:04:49.779 --> 00:04:52.779 regular visible or infrared satellites, they're 00:04:52.779 --> 00:04:55.259 great for tracking clouds, seeing the storm top. 00:04:55.459 --> 00:04:58.459 But they struggle. They can't really see through 00:04:58.459 --> 00:05:01.860 the dense clouds, especially at night when visible 00:05:01.860 --> 00:05:04.800 light is gone anyway. Ah. So they completely 00:05:04.800 --> 00:05:07.540 miss crucial details, stuff happening inside 00:05:07.540 --> 00:05:09.860 the storm structure, changes that can signal 00:05:10.079 --> 00:05:12.819 dangerous rapid intensification, you just can't 00:05:12.819 --> 00:05:14.600 see the inner workings with those conventional 00:05:14.600 --> 00:05:17.680 tools. Right, you only see the surface, the top 00:05:17.680 --> 00:05:20.300 layer. Precisely. And what's remarkable, what 00:05:20.300 --> 00:05:23.000 makes microwave data indispensable, is it gives 00:05:23.000 --> 00:05:27.000 us this unique capability. It's almost like an 00:05:27.000 --> 00:05:30.199 MRI scan of the atmosphere. An MRI, okay, I like 00:05:30.199 --> 00:05:32.600 that analogy. Yeah, like a doctor using an x 00:05:32.600 --> 00:05:35.740 -ray or MRI to see through PISHU to the organs 00:05:35.740 --> 00:05:37.939 inside. That's what microwave satellites do for 00:05:37.939 --> 00:05:40.660 us. They let forecasters literally peer through 00:05:40.660 --> 00:05:42.600 the obscuring clouds, through the heavy rain, 00:05:42.879 --> 00:05:45.519 right down near the surface. So they can see 00:05:45.519 --> 00:05:47.980 the storm's inner workings. Exactly. Where the 00:05:47.980 --> 00:05:50.360 heaviest rain is, the actual center location, 00:05:50.439 --> 00:05:52.560 and critically, its thermal and moisture structure. 00:05:52.740 --> 00:05:54.759 And they get these updates every few hours. It 00:05:54.759 --> 00:05:57.060 reveals what's happening inside, even when it's 00:05:57.060 --> 00:05:59.240 totally hidden by clouds. That's the x -ray vision 00:05:59.240 --> 00:06:01.300 you mentioned, seeing the internal heat, the 00:06:01.300 --> 00:06:04.120 moisture, stuff the normal pictures miss. It's 00:06:04.120 --> 00:06:05.899 the difference between seeing a car's paint job 00:06:06.160 --> 00:06:08.199 and understanding how the engine works. It's 00:06:08.199 --> 00:06:10.540 that level of insight. OK, so let's zero in on 00:06:10.540 --> 00:06:14.139 the instrument itself, the one going dark. SSMIS 00:06:14.139 --> 00:06:16.980 Special Sensor Microwave Imager Sounder. Break 00:06:16.980 --> 00:06:19.300 that down for us. OK, so it's a conical scanning 00:06:19.300 --> 00:06:22.860 passive microwave radiometer. Whoa, OK. Conical 00:06:22.860 --> 00:06:25.120 scanning? Yeah, imagine a sensor spinning like 00:06:25.120 --> 00:06:27.720 a lighthouse beam sweeps around, but it's looking 00:06:27.720 --> 00:06:31.300 down, sweeping this wide cone -shaped path below 00:06:31.300 --> 00:06:34.269 the satellite as it orbits. Got it. And passive. 00:06:34.410 --> 00:06:37.209 Passive means it's not sending out its own signal 00:06:37.209 --> 00:06:40.350 like radar does. It's just listening. Listening 00:06:40.350 --> 00:06:43.029 to the natural microwave energy that the Earth's 00:06:43.029 --> 00:06:45.110 surface and the atmosphere, including the storm, 00:06:45.589 --> 00:06:48.220 emit naturally. Different things emit different 00:06:48.220 --> 00:06:51.300 signals exactly water vapor rain ice particles 00:06:51.300 --> 00:06:54.100 They all have unique microwave signatures and 00:06:54.100 --> 00:06:56.180 this instrument picks them up across four main 00:06:56.180 --> 00:07:02.040 frequencies 19 .3 22 .2 37 .0 and 85 .5 gigahertz 00:07:02.040 --> 00:07:04.779 and it uses seven specific channels within those 00:07:04.779 --> 00:07:06.819 with different polarizations to like different 00:07:06.819 --> 00:07:08.959 channels on a radio tuned to different things 00:07:08.959 --> 00:07:11.040 kind of yeah Think of them as different wavelengths 00:07:11.040 --> 00:07:13.899 each tuned to see different aspects lower frequencies 00:07:13.899 --> 00:07:17.050 are good for heavy rain 22 .2 gigahertz is super 00:07:17.050 --> 00:07:19.790 sensitive to water vapor, and that 85 .5 gigahertz 00:07:19.790 --> 00:07:22.170 one. That gives incredible detail on ice scattering 00:07:22.170 --> 00:07:24.790 deep inside the storm core, really vital for 00:07:24.790 --> 00:07:26.490 structure. And this tech isn't brand new, right? 00:07:26.550 --> 00:07:28.350 You mentioned predecessors. All right, the first 00:07:28.350 --> 00:07:31.370 one, the SSMI, launched way back in June 1987, 00:07:31.589 --> 00:07:33.410 part of the Defense Meteorological Satellite 00:07:33.410 --> 00:07:36.730 Program, DMSP. So this capability, it's been 00:07:36.730 --> 00:07:39.779 foundational for decades. And that history, that 00:07:39.779 --> 00:07:42.399 long record, makes this loss even more significant. 00:07:42.800 --> 00:07:45.639 These DMST satellites with SSMIS, they orbit 00:07:45.639 --> 00:07:48.519 in a specific way. Yeah, near constant 53 degree 00:07:48.519 --> 00:07:51.720 angle, about 830 kilometers up. And this geometry 00:07:51.720 --> 00:07:54.259 lets them collect consistent data over a really 00:07:54.259 --> 00:07:57.360 wide path. about 1 ,400 kilometers across. So 00:07:57.360 --> 00:08:00.560 each pass covers a huge area. A huge area, giving 00:08:00.560 --> 00:08:03.160 comprehensive observations and the resolution, 00:08:03.259 --> 00:08:06.019 how much detail we get. It's 25 kilometers for 00:08:06.019 --> 00:08:08.079 things like total water vapor and sea surface 00:08:08.079 --> 00:08:10.339 wind and five brightness temperature channels. 00:08:10.569 --> 00:08:14.329 But for that crucial 85 .5 GHz channel, the one 00:08:14.329 --> 00:08:16.589 giving the sharpest internal view, it's even 00:08:16.589 --> 00:08:19.050 finer, 12 .5 km. Well, it aren't just numbers, 00:08:19.089 --> 00:08:20.949 are they? That defines the quality the detail 00:08:20.949 --> 00:08:23.050 forecasters have relied on. It's high definition 00:08:23.050 --> 00:08:25.329 storm diagnostics. Absolutely invaluable. So 00:08:25.329 --> 00:08:27.509 what do forecasters do with all this detailed 00:08:27.509 --> 00:08:31.029 penetrating data beyond just cool images? Oh, 00:08:31.230 --> 00:08:33.889 its contributions are indispensable. Huge for 00:08:33.889 --> 00:08:37.490 accuracy. For starters, it's critical for precisely 00:08:37.490 --> 00:08:40.440 fixing the storm center. pinpointing it exactly. 00:08:40.559 --> 00:08:42.539 That sounds simple, but you're saying it's not. 00:08:42.600 --> 00:08:44.919 It's notoriously difficult with just visible 00:08:44.919 --> 00:08:47.799 or infrared, especially if the center is hidden 00:08:47.799 --> 00:08:50.159 under a big cloud shield, which often happens. 00:08:50.580 --> 00:08:53.279 Like finding a needle in a haystack in the dark. 00:08:53.679 --> 00:08:55.320 Microwave lights up that needle. And getting 00:08:55.320 --> 00:08:58.320 that position right is crucial. Because it directly 00:08:58.320 --> 00:09:01.409 fights that. Butterfly effect we mentioned. Even 00:09:01.409 --> 00:09:04.409 small errors, maybe just tens of miles off initially 00:09:04.409 --> 00:09:07.210 for a weaker storm. They blow up over time. Dramatically. 00:09:07.470 --> 00:09:10.009 That tiny initial error, if you don't correct 00:09:10.009 --> 00:09:12.610 it with precise microwave data, it cascades into 00:09:12.610 --> 00:09:15.350 outsize forecast errors four or five days later. 00:09:15.830 --> 00:09:18.129 Big misses in the track forecast. But where the 00:09:18.129 --> 00:09:21.190 storm actually goes, that initial pinpoint matters 00:09:21.190 --> 00:09:23.889 that much. Immensely. It's like aiming that billiard 00:09:23.889 --> 00:09:25.950 ball off by a millimeter at the start. You're 00:09:25.950 --> 00:09:29.169 way off down the table. Without SSMIS, that initial 00:09:29.169 --> 00:09:31.690 millimeter error is much more likely. Wow. And 00:09:31.690 --> 00:09:35.509 it also powers some advanced tech AI stuff. Absolutely 00:09:35.509 --> 00:09:37.549 essential for that. One of the key tools for 00:09:37.549 --> 00:09:39.309 estimating storm intensity, especially when the 00:09:39.309 --> 00:09:42.269 Hurricane Hunters can't fly, is called DMINT, 00:09:43.090 --> 00:09:45.850 Deep Multispectral Intensity of Tropical Cyclones 00:09:45.850 --> 00:09:50.549 Estimator. DMINT sounds powerful. It is. It uses 00:09:50.549 --> 00:09:53.289 AI to analyze satellite data and predict storm 00:09:53.289 --> 00:09:56.179 strength. and it relies heavily, almost entirely, 00:09:56.419 --> 00:09:59.639 on the microwave data it ingests. Without SSMS, 00:09:59.899 --> 00:10:02.840 DMINT's ability to give accurate, timely intensity 00:10:02.840 --> 00:10:06.179 estimates severely hampered. Forecasters lose 00:10:06.179 --> 00:10:08.259 a key intensity tool. Leaving them less certain 00:10:08.259 --> 00:10:10.940 about the storm's true power. Exactly. And that 00:10:10.940 --> 00:10:12.960 directly increases the risk of what James Franklin, 00:10:13.340 --> 00:10:16.220 a retired NEC branch chief, famously called a 00:10:16.220 --> 00:10:19.240 sunrise surprise. A sunrise surprise? Sounds 00:10:19.240 --> 00:10:21.559 ominous. It is. It's that horrible moment seeing 00:10:21.559 --> 00:10:23.539 the first visible satellite pictures at dawn 00:10:23.539 --> 00:10:26.620 and realizing a storm got way stronger, way better 00:10:26.620 --> 00:10:29.080 organized overnight, maybe intensified dramatically. 00:10:29.179 --> 00:10:31.269 But you didn't see it happening. Because the 00:10:31.269 --> 00:10:33.990 key structural details were hidden from the infrared 00:10:33.990 --> 00:10:37.190 views overnight. Details only microwave data 00:10:37.190 --> 00:10:40.009 would have shown. It's finding out too late that 00:10:40.009 --> 00:10:42.409 the storm ramped up, giving people much less 00:10:42.409 --> 00:10:44.789 time to prepare for something far more dangerous. 00:10:44.909 --> 00:10:46.870 That's terrifying. And that scenario becomes 00:10:46.870 --> 00:10:49.970 more likely now. Much more likely. That microwave 00:10:49.970 --> 00:10:52.669 data was our safeguard against it. So the bottom 00:10:52.669 --> 00:10:55.740 line here, the immediate impact. DMSP, these 00:10:55.740 --> 00:10:59.100 three satellites with SSMIS, they provide roughly 00:10:59.100 --> 00:11:01.480 half of all the microwave scans forecasters get. 00:11:01.759 --> 00:11:04.419 Roughly half, yeah. Globally. And with the DoD 00:11:04.419 --> 00:11:07.139 pulling the plug on processing by June 30th, 00:11:07.440 --> 00:11:09.860 that vital data stream just goes dark. Poof. 00:11:10.059 --> 00:11:12.500 Gone. It's like a surgeon suddenly losing half 00:11:12.500 --> 00:11:15.240 their x -rays mid -operation, or a pilot losing 00:11:15.240 --> 00:11:17.200 half their instruments on approach. It's that 00:11:17.200 --> 00:11:19.139 level of challenge for hurricane forecasters 00:11:19.139 --> 00:11:22.100 now. This isn't just an inconvenience. The very 00:11:22.100 --> 00:11:24.639 data that helps spot those sudden terrifying 00:11:24.639 --> 00:11:27.360 bursts is rapid intensification. When a tropical 00:11:27.360 --> 00:11:29.840 storm explodes into a cat three or four overnight, 00:11:30.259 --> 00:11:32.899 that early warning capability is severely compromised. 00:11:33.230 --> 00:11:36.590 data availability sliced in half. It just greatly 00:11:36.590 --> 00:11:39.429 increases the odds of missing these events, underestimating 00:11:39.429 --> 00:11:42.190 storm strength, misplacing centers. It degrades 00:11:42.190 --> 00:11:45.750 forecast accuracy, period. It's a huge blow to 00:11:45.750 --> 00:11:48.529 hurricane preparedness, and it inherently puts 00:11:48.529 --> 00:11:50.610 more people at risk, right? Could cost people 00:11:50.610 --> 00:11:53.230 more money, dealing with worse forecasts? Absolutely. 00:11:53.429 --> 00:11:55.850 More risk, potentially higher costs for everyone 00:11:55.850 --> 00:11:59.019 involved, individuals, communities. Dealing with 00:11:59.019 --> 00:12:01.480 degraded forecasts has real consequences. And 00:12:01.480 --> 00:12:03.700 this isn't just about hurricanes, is it? The 00:12:03.700 --> 00:12:06.440 impact spread wider. Much wider. This data determination 00:12:06.440 --> 00:12:08.860 sends ripples everywhere, including critical 00:12:08.860 --> 00:12:11.179 climate monitoring, especially for sea ice in 00:12:11.179 --> 00:12:13.179 the polar regions. Sea ice? How does microwave 00:12:13.179 --> 00:12:16.519 data help there? Well, SSMIS data is fundamental. 00:12:16.740 --> 00:12:18.679 It's a core input for key data sets from the 00:12:18.679 --> 00:12:21.659 National Snow and Ice Data Center, NSIDC, run 00:12:21.659 --> 00:12:23.980 by the National Oceanic and Atmospheric Administration. 00:12:24.169 --> 00:12:27.009 Things like their near real -time sea ice concentration 00:12:27.009 --> 00:12:30.330 records, the overall sea ice index, crucial for 00:12:30.330 --> 00:12:32.470 tracking climate change, Arctic navigation. And 00:12:32.470 --> 00:12:36.009 those products rely on SSMIS. Heavily. And the 00:12:36.009 --> 00:12:39.149 immediate unsettling consequence. Those specific 00:12:39.149 --> 00:12:41.990 products just stop processing after June 30th, 00:12:42.070 --> 00:12:46.129 2025. Users, scientists, shippers, northern communities, 00:12:46.230 --> 00:12:48.539 they have to anticipate a data gap. Like losing 00:12:48.539 --> 00:12:51.039 a key chapter in the climate story makes it harder 00:12:51.039 --> 00:12:53.259 to see the whole picture. It's a major disruption 00:12:53.259 --> 00:12:55.980 for climate science. And navigating it, well, 00:12:56.279 --> 00:12:58.539 there's this urgent scramble to find and integrate 00:12:58.539 --> 00:13:02.860 alternatives for products built purely on SSMIS 00:13:02.860 --> 00:13:06.779 like sea ice concentration, OSI, 401D, and the 00:13:06.779 --> 00:13:10.220 interim record OSI 430A users just have to switch. 00:13:10.570 --> 00:13:12.970 find new products based on different data. Which 00:13:12.970 --> 00:13:15.889 isn't simple, right? Requires validation, introduces 00:13:15.889 --> 00:13:18.730 uncertainty. Exactly. It takes time, resources, 00:13:19.190 --> 00:13:21.049 not a clean swap. But some products are okay. 00:13:21.049 --> 00:13:23.070 You mentioned multi -sensor ones. Yeah, there's 00:13:23.070 --> 00:13:24.850 a bit of a silver lining there. Things like Sea 00:13:24.850 --> 00:13:28.750 Ice Edge, Type, and Drift OSI 402D, 403, 405D, 00:13:28.850 --> 00:13:31.210 they use multiple sensors already. And another 00:13:31.210 --> 00:13:33.649 instrument, AMSR2, is their primary source, so 00:13:33.649 --> 00:13:34.950 they're more resilient. They'll keep working 00:13:34.950 --> 00:13:37.529 without SSMIS. Okay, so they had a backup plan, 00:13:37.669 --> 00:13:39.919 essentially. That's fortunate. incredibly fortunate 00:13:39.919 --> 00:13:42.639 for those specific applications. And other products 00:13:42.639 --> 00:13:45.919 like sea ice emissivity and L2C ice concentration, 00:13:46.220 --> 00:13:49.240 they're either actively migrating to use AMSR2 00:13:49.240 --> 00:13:52.539 data now, or they already offer AMSR2 versions 00:13:52.539 --> 00:13:54.799 that users are being urged to switch to. So there's 00:13:54.799 --> 00:13:58.220 a push to adapt. A huge push. And what's really 00:13:58.220 --> 00:14:00.059 critical for the long -term climate picture is 00:14:00.059 --> 00:14:02.740 things like the Sea Ice Index, OSI 420. They're 00:14:02.740 --> 00:14:05.440 undergoing complex changes to integrate AMSR2. 00:14:05.860 --> 00:14:08.700 The absolute goal is a continuous time series, 00:14:09.200 --> 00:14:12.100 seamlessly building on the old SSMIS data. So 00:14:12.100 --> 00:14:14.120 we don't lose track of the long -term trends. 00:14:14.340 --> 00:14:17.100 Precisely. That continuity is vital. It's a massive 00:14:17.100 --> 00:14:19.840 effort, but essential to preserve decades of 00:14:19.840 --> 00:14:21.860 climate records. OK, so we've covered the practical 00:14:21.860 --> 00:14:24.539 impacts, hurricanes, climate. Let's really dive 00:14:24.539 --> 00:14:27.379 into the science now. data assimilation, why 00:14:27.379 --> 00:14:29.860 this specific data loss hurts so badly. Right. 00:14:30.000 --> 00:14:31.940 At the very foundation of all weather prediction 00:14:31.940 --> 00:14:34.600 models, these incredibly complex computer programs 00:14:34.600 --> 00:14:37.519 generating forecasts is this absolute reliance 00:14:37.519 --> 00:14:39.840 on accurate initial conditions. Starting point. 00:14:40.360 --> 00:14:42.879 Exactly. Precise data on the atmosphere state 00:14:42.879 --> 00:14:45.519 right now, especially moisture. Think of it like 00:14:45.519 --> 00:14:48.029 setting up dominoes. if those first few are off 00:14:48.029 --> 00:14:50.950 even slightly. The whole chain falls apart unpredictably. 00:14:51.049 --> 00:14:53.629 Perfectly put. And this is where our conventional 00:14:53.629 --> 00:14:56.009 observations weather balloons launched mostly 00:14:56.009 --> 00:14:59.190 over land twice a day. Surface stations, also 00:14:59.190 --> 00:15:01.620 mostly land -based, just fall short. especially 00:15:01.620 --> 00:15:03.740 for fast -changing storms over the ocean. They're 00:15:03.740 --> 00:15:06.940 too sparse, too infrequent for hurricanes. Entirely 00:15:06.940 --> 00:15:09.240 inadequate for sampling those dynamic, moisture 00:15:09.240 --> 00:15:11.919 -rich environments over the vast oceans. And 00:15:11.919 --> 00:15:14.740 even tiny uncertainties in that initial humidity 00:15:14.740 --> 00:15:16.720 field, which is super hard to measure globally, 00:15:17.139 --> 00:15:19.480 they can have a huge detrimental impact later 00:15:19.480 --> 00:15:22.340 on in the forecast. Errors grow exponentially. 00:15:22.480 --> 00:15:24.600 Which is where satellites become absolutely essential. 00:15:25.000 --> 00:15:27.570 Irreplaceable, even. Totally. Unlike fixed points, 00:15:27.929 --> 00:15:30.409 satellites sweep across the globe, filling those 00:15:30.409 --> 00:15:33.169 massive ocean gaps, providing spatially dense, 00:15:33.370 --> 00:15:35.850 repeatable data where we had almost nothing before. 00:15:36.129 --> 00:15:39.049 From scattered snapshots to a continuous global 00:15:39.049 --> 00:15:42.029 video feed. A great way to think about it. Now, 00:15:42.269 --> 00:15:45.210 getting all this indirect satellite data, not 00:15:45.210 --> 00:15:47.690 direct temperature or pressure, but microwave 00:15:47.690 --> 00:15:50.590 signals into the models, that requires sophisticated 00:15:50.590 --> 00:15:54.320 techniques. Data simulation. Specifically, things 00:15:54.320 --> 00:15:57.360 like 3DVAR, three -dimensional variational systems. 00:15:57.559 --> 00:16:00.120 3DVAR. Okay, break that down. Imagine the atmosphere 00:16:00.120 --> 00:16:03.379 is a giant complex puzzle. Our models predict 00:16:03.379 --> 00:16:06.259 how the pieces move. Traditional observations 00:16:06.259 --> 00:16:08.820 give us some pieces, but there are huge gaps, 00:16:08.860 --> 00:16:11.500 especially over oceans. Satellite data floods 00:16:11.500 --> 00:16:14.539 us with new clues, often indirect clues like 00:16:14.539 --> 00:16:17.039 microwave brightness temperatures. So 3DVAR is 00:16:17.039 --> 00:16:19.519 how we fit those clues into the puzzle. Exactly. 00:16:19.600 --> 00:16:22.029 It's like a super smart digital detective. It 00:16:22.029 --> 00:16:23.870 takes all observations, balloons, satellites, 00:16:24.090 --> 00:16:25.649 everything, and mathematically finds the single 00:16:25.649 --> 00:16:27.990 most consistent and most probable picture of 00:16:27.990 --> 00:16:30.629 the atmosphere right now. It optimizes the initial 00:16:30.629 --> 00:16:33.049 conditions, gets the puzzle started as accurately 00:16:33.049 --> 00:16:34.809 as possible. So the rest of the prediction works 00:16:34.809 --> 00:16:36.789 better. And there are different ways to feed 00:16:36.789 --> 00:16:39.370 SSMIS data into this. Two main ways, yeah. First 00:16:39.370 --> 00:16:42.789 is assimilating retrieved products, or RV. RV, 00:16:43.110 --> 00:16:45.669 what's that? It's computationally faster. You 00:16:45.669 --> 00:16:48.110 first convert the raw satellite signals into 00:16:48.110 --> 00:16:50.509 geophysical things like total water vapor or 00:16:50.509 --> 00:16:53.549 surface wind using statistical algorithms. Then 00:16:53.549 --> 00:16:56.549 you feed those into the model. Quicker, but maybe 00:16:56.549 --> 00:16:59.529 less accurate. Potentially. Because those statistical 00:16:59.529 --> 00:17:02.110 conversions, those retrievals, they might have 00:17:02.110 --> 00:17:05.670 errors or biases if the statistics aren't perfect 00:17:05.670 --> 00:17:08.630 or if other info used in the process is flawed. 00:17:09.029 --> 00:17:10.950 It's an estimation. OK, so what's the other way? 00:17:11.160 --> 00:17:13.720 The second method is directly assimilating the 00:17:13.720 --> 00:17:15.900 raw measurements, the brightness temperatures, 00:17:16.160 --> 00:17:20.319 or TB. TB raw data. Right. This is more computationally 00:17:20.319 --> 00:17:22.819 intensive but more physically robust. You feed 00:17:22.819 --> 00:17:24.880 the raw microwave signals directly into the model 00:17:24.880 --> 00:17:28.259 using a complex radiative transfer model. This 00:17:28.259 --> 00:17:30.400 simulates how microwaves actually travel through 00:17:30.400 --> 00:17:32.380 the atmosphere. So it's based more on physics 00:17:32.380 --> 00:17:35.700 than statistics. Exactly. Less dependent on potentially 00:17:35.700 --> 00:17:38.779 iffy stats. Better at reducing systematic biases. 00:17:39.289 --> 00:17:42.049 It takes more computer power, but often gives 00:17:42.049 --> 00:17:44.369 a pure, more accurate starting point because 00:17:44.369 --> 00:17:46.710 it uses the fundamental physics. More complex, 00:17:46.950 --> 00:17:49.349 but potentially better results. Generally, yes. 00:17:49.630 --> 00:17:51.549 And the math works out well enough to make it 00:17:51.549 --> 00:17:53.609 feasible. Okay, to really hammer home why this 00:17:53.609 --> 00:17:56.039 matters, you mentioned a case study. Hurricane 00:17:56.039 --> 00:17:59.259 Danny. Yeah, Hurricane Danny back in July 1997. 00:17:59.980 --> 00:18:01.960 Fascinating storm formed from a non -tropical 00:18:01.960 --> 00:18:04.660 system intensified in the Gulf. A perfect real 00:18:04.660 --> 00:18:07.380 -world test bed for seeing the impact of SSMI 00:18:07.380 --> 00:18:09.960 data assimilation. So what happened in the baseline? 00:18:10.420 --> 00:18:13.599 The forecasts without the microwave data. The 00:18:13.599 --> 00:18:16.940 control case. The results were stark. Pretty 00:18:16.940 --> 00:18:19.220 alarming, actually. The initial conditions from 00:18:19.220 --> 00:18:22.460 standard systems like NOAAs, GDAs, they severely 00:18:22.460 --> 00:18:24.690 lacked moisture over the Gulf. And the storm's 00:18:24.690 --> 00:18:27.369 initial spin, the low level circulation, was 00:18:27.369 --> 00:18:29.730 way too weak. Which led to? Big underestimates 00:18:29.730 --> 00:18:32.029 of intensity. For example, the model's central 00:18:32.029 --> 00:18:35.630 pressure for Danny was 19 .5 hectopascals higher 00:18:35.630 --> 00:18:38.849 than observed after 48 hours. A huge miss. Plus 00:18:38.849 --> 00:18:40.809 the model needed at least eight hours of spin 00:18:40.809 --> 00:18:42.950 up time before it even started raining realistically. 00:18:43.190 --> 00:18:45.109 It was starting cold. Eight hours just to get 00:18:45.109 --> 00:18:47.829 going properly. That's a long delay. A very long 00:18:47.829 --> 00:18:50.910 delay. But here's where the SSMI data made a 00:18:50.910 --> 00:18:53.529 transformative difference. When they assimilated, 00:18:53.680 --> 00:18:56.880 Either the retrieved products, RV, or the raw 00:18:56.880 --> 00:19:00.019 brightness temps, TB. What changed? Instantly, 00:19:00.259 --> 00:19:03.019 almost. The atmospheric moisture across most 00:19:03.019 --> 00:19:05.619 of the Gulf dramatically increased in the model 00:19:05.619 --> 00:19:08.119 starting point. The low level circulation got 00:19:08.119 --> 00:19:11.140 much stronger, much more realistic. The inflow 00:19:11.140 --> 00:19:13.420 feeding the storm improved. It gave the model 00:19:13.420 --> 00:19:15.380 the right ingredients from the start. Exactly. 00:19:15.759 --> 00:19:18.240 And crucially, that spin up time. It plummeted. 00:19:18.500 --> 00:19:21.240 With SSMI data in, the model started producing 00:19:21.240 --> 00:19:23.299 light rain within just one or two hours. Wow, 00:19:23.319 --> 00:19:26.259 from eight hours down to one or two. That's huge 00:19:26.259 --> 00:19:29.240 for getting timely rain forecasts out. Massive. 00:19:29.400 --> 00:19:32.140 And the intensity forecasts. They got incredibly 00:19:32.140 --> 00:19:34.240 better, much more accurate capturing the storm's 00:19:34.240 --> 00:19:36.970 deepening pressure. The raw data. TB approach 00:19:36.970 --> 00:19:39.269 got the final pressure within one hectopascal 00:19:39.269 --> 00:19:42.809 of reality after 48 hours. The retrieved data, 00:19:42.990 --> 00:19:46.170 RV, was 8 HPA deeper than observed, still vastly 00:19:46.170 --> 00:19:49.309 better than the control's 19 .5 HPA error. So 00:19:49.309 --> 00:19:51.569 both methods were dramatically better than no 00:19:51.569 --> 00:19:54.029 data at all. A revolution, really. Absolutely. 00:19:54.410 --> 00:19:56.190 And the visual structure in the simulations. 00:19:56.809 --> 00:20:00.450 Night and day. With SSMI, the models produced 00:20:00.700 --> 00:20:03.660 clear hurricane eyes, distinct eye walls, spiral 00:20:03.660 --> 00:20:06.180 bands, which is totally missing in the control 00:20:06.180 --> 00:20:08.700 run. Like seeing an ultrasound versus just guessing? 00:20:09.059 --> 00:20:11.740 A perfect analogy. Forecasters could suddenly 00:20:11.740 --> 00:20:14.180 see the storm's internal evolution much more 00:20:14.180 --> 00:20:16.400 clearly. Now what about the computational cost? 00:20:16.839 --> 00:20:19.519 All that raw data processing sounds expensive. 00:20:19.640 --> 00:20:21.759 You'd think so, but the study actually showed 00:20:21.759 --> 00:20:24.180 that while the raw TB method took about twice 00:20:24.180 --> 00:20:27.500 the CPU time as the RV method, the total assimilation 00:20:27.500 --> 00:20:30.480 time was still only about 3 .5 minutes. compared 00:20:30.480 --> 00:20:33.039 to the 157 hours it took to run the 48 -hour 00:20:33.039 --> 00:20:35.980 forecast itself. Negligible. So it's totally 00:20:35.980 --> 00:20:38.400 affordable, computationally speaking. A tiny 00:20:38.400 --> 00:20:40.480 investment for massive forecast improvement. 00:20:40.799 --> 00:20:42.700 Completely affordable and highly beneficial. 00:20:42.900 --> 00:20:45.259 It pays for itself many times over in improved 00:20:45.259 --> 00:20:47.799 warning. But it wasn't a magic bullet for everything, 00:20:47.819 --> 00:20:50.619 right? The study noted some limitations. True. 00:20:51.200 --> 00:20:53.619 While intensity and structure improved dramatically, 00:20:54.059 --> 00:20:56.420 the impact on the storm track forecast was surprisingly 00:20:56.420 --> 00:21:00.190 small. All the simulations, even with SSMI data, 00:21:00.549 --> 00:21:03.549 had this consistent leftward track bias. They 00:21:03.549 --> 00:21:05.369 predicted Danny to go slightly left of where 00:21:05.369 --> 00:21:07.230 it actually went. And that matched the official 00:21:07.230 --> 00:21:09.690 forecasts at the time. Yeah, suggesting a deeper 00:21:09.690 --> 00:21:13.049 issue. Probably inaccuracies in the initial steering 00:21:13.049 --> 00:21:15.369 winds higher up in the atmosphere. Stuff that 00:21:15.369 --> 00:21:17.650 SSMI doesn't directly measure well. Exactly. 00:21:17.869 --> 00:21:19.630 It points to the need for other instruments, 00:21:19.869 --> 00:21:21.650 like LiR, that can measure those steering winds 00:21:21.650 --> 00:21:24.359 directly. You need complementary data. And even 00:21:24.359 --> 00:21:26.819 with SSMI, the model still struggled to capture 00:21:26.819 --> 00:21:29.359 every cloud band perfectly, or subtle things 00:21:29.359 --> 00:21:31.980 like cirrus clouds, no single instrument sees 00:21:31.980 --> 00:21:34.140 everything. It's always about combining perspectives. 00:21:34.319 --> 00:21:37.599 Always. But the key takeaway from Danny is undeniable. 00:21:38.359 --> 00:21:41.539 Assimilating that SSMI data had a strong, unequivocally 00:21:41.539 --> 00:21:44.140 positive impact, especially on intensity and 00:21:44.140 --> 00:21:46.839 structure. While the two methods, RV versus TB, 00:21:46.880 --> 00:21:49.500 had differences, both were vastly superior to 00:21:49.500 --> 00:21:52.079 having no data. It proved this microwave data 00:21:52.079 --> 00:21:55.130 is foundational. Foundational, not just incremental. 00:21:55.890 --> 00:21:58.069 Okay, so laying all that out, knowing its value, 00:21:59.490 --> 00:22:02.710 what's the path forward now? We have this immediate 00:22:02.710 --> 00:22:05.869 lasting data gap. Yeah, and it casts a long shadow 00:22:05.869 --> 00:22:08.470 over this hurricane season and beyond. Losing 00:22:08.470 --> 00:22:11.009 SSMIS means losing half our microwave instruments. 00:22:11.690 --> 00:22:14.049 It dangerously increases the odds of missing 00:22:14.049 --> 00:22:16.890 rapid intensification, underestimating intensity, 00:22:17.329 --> 00:22:19.589 misplacing storms. Putting more people at risk, 00:22:19.730 --> 00:22:22.150 increasing potential costs. Plain and simple. 00:22:22.369 --> 00:22:24.450 But the community isn't just sitting back. They're 00:22:24.450 --> 00:22:27.089 urgently searching for replacements. The main 00:22:27.089 --> 00:22:29.890 contender right now is AMSR2 Advanced Microwave 00:22:29.890 --> 00:22:33.150 Scanning Radiometer 2. AMSR2, does it offer improvements? 00:22:33.309 --> 00:22:36.250 It does. Better resolution than SSMIS, so potentially 00:22:36.250 --> 00:22:38.490 more detail. That's good. Out there is a catch. 00:22:38.710 --> 00:22:41.309 There is. AMSR2 data is still considered prototype. 00:22:41.490 --> 00:22:43.890 It's undergoing testing. And crucially, it hasn't 00:22:43.890 --> 00:22:46.710 been fully intercalibrated with the older SSMIS 00:22:46.710 --> 00:22:48.789 data yet. Meaning the measurements might not 00:22:48.789 --> 00:22:52.039 line up perfectly. Exactly. Users might see differences, 00:22:52.079 --> 00:22:55.680 say in CI610, comparing AMSR2 to the historical 00:22:55.680 --> 00:22:59.039 SSMIS record. partly because its higher resolution 00:22:59.039 --> 00:23:01.839 sees edges better. So while it's promising, it's 00:23:01.839 --> 00:23:04.099 not a seamless swap yet. It's a work in progress. 00:23:04.200 --> 00:23:06.079 And progress takes time we don't necessarily 00:23:06.079 --> 00:23:08.619 have right now. What about other options? There's 00:23:08.619 --> 00:23:11.740 the weather system follow on microwave, WSFM, 00:23:11.880 --> 00:23:16.259 designed specifically as the successor to DMSPS 00:23:16.259 --> 00:23:19.480 SMIS. It launched successfully in April 2024. 00:23:19.819 --> 00:23:22.039 Fantastic. So that should fill the gap. It should. 00:23:22.400 --> 00:23:24.990 But here's the frustrating part. The data from 00:23:24.990 --> 00:23:28.250 WSF -M is not currently available to forecasters. 00:23:28.430 --> 00:23:30.150 Wait, it's up there, but we can't use the data. 00:23:30.190 --> 00:23:32.930 Correct. It's uncertain if or when access will 00:23:32.930 --> 00:23:35.170 be permitted. Bureaucratic hurdles, security 00:23:35.170 --> 00:23:38.049 protocols, complex reasons. But the bottom line 00:23:38.049 --> 00:23:41.210 for forecasters, a capable satellite designed 00:23:41.210 --> 00:23:43.650 for this exact purpose is flying blindfolded 00:23:43.650 --> 00:23:45.990 right now. The data isn't flowing. That's incredibly 00:23:45.990 --> 00:23:48.269 frustrating. So connecting this all back, the 00:23:48.269 --> 00:23:51.170 broader stakes for society are huge. Undeniable. 00:23:51.769 --> 00:23:54.640 This data loss... puts more people at risk. It 00:23:54.640 --> 00:23:57.099 could lead to higher costs, unnecessary evacuations, 00:23:57.259 --> 00:24:00.079 or worse, failing to evacuate when needed. The 00:24:00.079 --> 00:24:02.740 ripples hit emergency services, insurance, supply 00:24:02.740 --> 00:24:05.420 chains, daily lives. It really underscores how 00:24:05.420 --> 00:24:08.019 much we rely on this infrastructure. On a slightly 00:24:08.019 --> 00:24:09.799 brighter note, anything happening right now? 00:24:10.420 --> 00:24:13.859 A tiny fleeting positive. Current models show 00:24:13.859 --> 00:24:15.640 the Atlantic staying quiet through the end of 00:24:15.640 --> 00:24:18.380 June. No tropical development expected for the 00:24:18.380 --> 00:24:21.660 next week or so. Okay, so a small window of reprieve 00:24:21.660 --> 00:24:24.140 as this data stream shuts down. A small window, 00:24:24.400 --> 00:24:26.859 but just a brief calm before what could be a 00:24:26.859 --> 00:24:29.079 very challenging, very unpredictable season, 00:24:29.579 --> 00:24:32.380 now with fewer tools in the toolbox. So to wrap 00:24:32.380 --> 00:24:34.579 up our Meteorology Matters discussion today, 00:24:34.940 --> 00:24:36.920 we've looked at the sudden, really impactful 00:24:36.920 --> 00:24:40.319 loss of critical SSMIS data. An indispensable 00:24:40.319 --> 00:24:43.220 eye in the sky, now dark. We've detailed the 00:24:43.220 --> 00:24:45.099 profound implications for hurricane intensity 00:24:45.099 --> 00:24:47.940 forecasts for vital sea ice monitoring that tracks 00:24:47.940 --> 00:24:50.200 climate change. And we've looked at the urgent 00:24:50.200 --> 00:24:52.779 ongoing scramble to find, test, and integrate 00:24:52.779 --> 00:24:55.059 alternatives to fill this sudden gaping hole 00:24:55.059 --> 00:24:57.529 in our observations. And this whole situation, 00:24:57.769 --> 00:24:59.769 it really makes you think, doesn't it? It raises 00:24:59.769 --> 00:25:02.349 a question for you, our listener. How often do 00:25:02.349 --> 00:25:05.470 we really stop to consider the complex, often 00:25:05.470 --> 00:25:08.609 invisible scientific web that supports our safety, 00:25:08.849 --> 00:25:12.369 our economy, just our daily lives? Yeah. What 00:25:12.369 --> 00:25:15.529 other crucial unseen data sources are we just 00:25:15.529 --> 00:25:18.910 assuming will always be there? Data that, like 00:25:18.910 --> 00:25:22.289 SSMIS, could just vanish overnight. And what's 00:25:22.289 --> 00:25:25.940 the real cost? of unquantifiable when these critical 00:25:25.940 --> 00:25:28.759 systems are disrupted. Something to ponder. Definitely 00:25:28.759 --> 00:25:30.420 something to ponder. Now, if you want to stay 00:25:30.420 --> 00:25:33.339 right on the cutting edge of meteorology, hurricane 00:25:33.339 --> 00:25:35.619 tracking, and keep getting clarity on vital topics 00:25:35.619 --> 00:25:38.200 like this, you absolutely need to follow meteorologist 00:25:38.200 --> 00:25:40.839 Rob Jones. You can find him on Instagram to search. 00:25:41.859 --> 00:25:45.599 He's also on TikTok as TVMeteorologist. And definitely 00:25:45.599 --> 00:25:47.720 check out his YouTube channel, search Rob Jones 00:25:47.720 --> 00:25:49.839 Hurricane. That's where you'll also find the 00:25:49.839 --> 00:25:51.960 Meteorology Matters podcast playlist with all 00:25:51.960 --> 00:25:54.460 our discussions. Excellent resources. Well, thank 00:25:54.460 --> 00:25:56.440 you for joining us on this Meteorology Matters 00:25:56.440 --> 00:25:58.480 deep dive. We look forward to exploring more 00:25:58.480 --> 00:25:59.819 essential knowledge with you next time.