 Excellent, thank you. That was a really exciting talk. I've been saying for years, the arctica eslendica is coming. And now it's nice to see some beta. OK, so I'm going to talk to you today about the low latitude perspective on this pretty much same topic. So a couple of years ago, I started sort of, I actually started looking into what we have in terms of paleo data that can address Atlantic multi-decadal climate variability. And really, the literature seemed to be asking one question. And that is what's up here, and is there a concentration of variants at multi-decadal bands in the climate of the Atlantic region that extends above the expected red noise background? And extends back through time. So we have here the classic sea surface temperature anomalies in the northern hemisphere in the Atlantic, detrended. And the question is, what is happening back here? And so the paper that came out of this actually published last year. And essentially in that, I do a detailed review of what we know. I sum it up in this single slide. Essentially, it seems like the consensus is that there has been significant multi-decadal variability back to the 1700s. But prior to the 1700s, different people say different things. Some people say it doesn't exist. Some people say it's intermittent. Some people say it's persistent. And in large part, it seems to me that the reason for all of these different answers is that we're looking at different proxies that are actually responding to different variables. So we're actually looking at apples and oranges. In many cases, it's a single record that people are making their statements about. And so those local signals may or may not capture the hemispheric-wide northern Atlantic sea surface temperature changes. And so my recommendation for the future after doing this whole review was essentially we need to be sure of what we are reconstructing. So rather than trying to use somewhat tele-connected variables, say tree rings to reconstruct ocean temperature, which while there is a response, there can be also many other variables that impact those tree rings, and those trees are not living in the ocean. So I'll respect to Edna. But then in fact, we should be using each variable for what it's best at. So tree rings for looking at terrestrial temperatures and droughts and marine proxies for looking at, say, sea surface temperatures. And then we can actually reconstruct variables that we can begin to tease out what's going on. OK, so that's my little soapbox. So I actually did a tropical reconstruction. I tend to be a low-latitude kind of gal. And as we've seen, a similar map. This is from Amy Clement's recent paper. Ah, I keep messing this up, where we have two centers of action, the high latitudes and the low latitudes. There seems to probably, I think, be some different dynamics going on in these two different regions. So I think it's interesting to get the higher latitude perspective from records like the Arctic Islandica. And then we have similarly high resolution, annual resolution proxies in the Caribbean and the tropical Atlantic to get the low-latitude perspective. And the low-latitudes, sea surface temperatures are important in part because of their connection to hurricane intensities and our frequencies and precipitation anomalies in other areas. So let's look at what I found when I tried to reconstruct all the variables that we had. OK, so I basically said, OK, we have to have some data selection criteria. So ocean temperature selective sensitive paleoproxies that's carbonate strontium calcium, magnesium oxygen 18, which is, of course, the dual signal. And then there have been some recent papers connecting coral growth with sea surface temperature. It's more of a local connection. So you have to do a local calibration bit of ad hoc. I personally prefer the chemistry because that's determined by the thermodynamics of crystallization. But the growth rate records calibrate well. So another important point of trying to reconstruct multi-decadal and decadal scale variabilities, you certainly have to have high resolution, well-dated proxies. And so that does limit us from sediment cores in most cases. So I made a requirement of less than five years per sample. I expanded it from one year per sample just because that helped me get more data. And I think that if you have a network of proxies, you can get away with a little bit lower resolution and still get reasonable certainty in your decadal variability. So the other thing was actually I made a limit of a location of south of 22 degrees north. And that's essentially based on this record of the sea surface temperature, which from Goldenberg et al. 2001. And you can see that this is the first DOF of global sea surface temperature after removing the trend and ENSO. And you get this pattern. And there's clearly this demarcation between the Gulf of Mexico and the Caribbean and the tropical Atlantic. There are several proxy records from up in here, but it's not highly correlated to the AMO index or multi-decadal sea surface temperature variability that we're trying to target. So I did not use those records specifically. So unfortunately, there wasn't a lot of data. So that was the first thing I learned from this and a take-home message is that I was able to get four sites, six records, done. And so that makes this less certain. The more data we can get, the more we can reduce our signal to noise ratios. And with payload data, you've always got quite a bit of uncertainty. So the one thing I did that was somewhat unique was that I used calibrations from the original publication or a species-specific calibration. So this is basically taking the paleoclimate theory of how do these proxies work and applying them rather than just fitting variance to a target, which is often done, but I think it can mess up some of your signal to noise ratios. So each record is an independent source of data, even though some of them come from the same site. And hopefully, the different noise, the different biases will average out. Ideally, more data would help that. So I created a target calibration. So I took all of the four sites I had. And instrumental sea surface temperatures averaged them up and got the green curve there. And it compares fairly well in terms of decadal variability. Certainly, there's some higher frequency stuff going on, higher amplitude. But you expect that when you're only dealing with four sites rather than a large regional average. So comparing my compilation of all of the records, that's the purple line with the air bars in light purple. And then that compared to my target. And it gave me a standard area of reconstruction of 0.34 degrees. Propagated that into the full reconstruction. Just want to point out that back here, we're down to one site. I no longer trust it. This is sort of back to 1360 has the full reconstruction. And so we see a centennial scale variability. That's obviously going to have a big effect on any time series analysis. So I actually sort of assessed that and pulled it out. So the in gray is the time series analysis of that first principal component of the data, that white curve that we saw just a minute ago. And you can see how that really does contain that centennial variability in the spectrum. So we subtract that first principal component and repeat the analysis. Of course, your centennial scale variability goes away. But our multi-decale scale variability remains. Interestingly, we have also variability at 35 and 16 years. I kind of think those are harmonics. But that I haven't done a lot of work on it. So I try to make that go away. I try to make that peak go away. One way I can do that is alter the noise floor and try to see if that's a spurious peak. Can't make it go away. Another thing I do is I actually cut the time series. Before 1850, we have the instrumental record, which has that very strong sort of 60-year apparent oscillation, whether it really is a whole other ball of wax. And so I cut that out to see if that was influencing the record, the strong 60-year period. And I got the gray line here. So you can see that that approximately 60-year period is pretty robust throughout the record. So that's sort of an interesting result. We've got this nice reconstruction of the tropical Atlantic Caribbean region back through 1360. And it seems to have a robust approximately 60-year sort of band of increased variance. I tried various iterations. So you can see the little dashed bars there are when I left different proxies out. So I would pretty much only trust this data as far as that darker purple sort of airbar band. Within that, I'm fairly sure we got it right. But anything finer than that, I would say there's still too much uncertainty. So comparing it with other AMO reconstructions is interesting. Remember, mine is simply a tropical. So to some extent, I don't expect it to perfectly match anything that has been tuned to the basin as a whole. However, the complete lack of correlation is rather disturbing. The gray bar here is a record from Gray et al. based on tree rings from North America, Scandinavia, and the Mediterranean. This is often cited as the AMO record from Paleo, which I find rather interesting. This one is from Mann et al. 2009. And he pulled that out from a spatial reconstruction based on a lot of multiple proxies. But in the text, he also notes that it probably has some admixture of PDO in it. So there's probably Pacific influence in it. The thing to note is they all pretty much agree on the multi-decadal pattern during the instrumental record. And then they have pretty much no relationship after that. So I would actually conclude from that is that we really are not there yet. We're still working on it in terms of what the history of sea surface temperatures in this region really are. Another thing that I've been working on is the Pages 2K project. And here, a group of us, we're working with high-resolution proxies. We actually did reconstructions for particular grid boxes and the tropics, each of these here. So I'll focus on the West Atlantic. Note that we used proxies from areas that are not strongly impacted by multi-decadal variability in the Atlantic. So to some extent, there's an issue there in the comparison. We used a composite plus scale methodology and we similarly did a leaf one-out iteration. All of those iterations are graphed here. The orange to red color indicates how well the data validated from our calibration interval, or separate from our calibration interval. And so you can see it actually didn't, it calibrated okay, but then it didn't validate very well. So we don't have a heck of a lot of confidence in this middle part and I would say we have not that much confidence in this one either. It just happened to be made up of a single data site that did validate well. So I still would like to see multiple records coming in before I trust it. So the take home message is when I put our reconstruction from the larger work that I did with the large Pages2K group, we have similarly lack of correlation. So we're still working on it. I wanted to highlight a project led by Casey Sanger and Mike Evans still related to the Pages2K project where they're trying to get at high latitude perspective. And let's see, what do I wanna say about this? I don't wanna go too fast. But basically they have lower resolution proxies, but similarly they extend over the last 2000 years and they're trying to use proxy surrogate reconstruction. So using information from climate models, essentially trying to match the model state to the paleo data, finding a best fit and then using that to help inform the interpretation of the paleo data. And the aspect of this that I think is really interesting is really this trying to merge or to utilize the physics in a climate model to help interpret and deal with the uncertainty in the paleo records. So this is actually the Pages2K trans-regional projects. You can look it up online. You can actually contact Casey and try to join. It's sort of a, they're trying to be a community-wide effort and then the other thing that I think is really exciting is the high-latitude bivalve records that we just saw a talk on. One last slide, I wanted to plug a meeting that I'm organizing along with several colleagues on trying to connect modern and paleo observations and modeling and work on Atlantic Merdionnel overturning circulation, so one of the potential drivers of this sea surface temperature anomaly. So we'll be meeting in May in Boulder and you can go online. It's a US Clivar Pages sponsored meeting and yeah, the registration will open up at the end of this month, so go look at that. So in summary, we need data from the high-latitudes and it's coming. I've reconstructed what we've got from the low-latitudes and it seems to show a persistent multi-decadal signal in the tropics, at least back to 1360. The, a clear history of the North Atlantic multi-decadal variability is still out there and hopefully getting closer, but I would argue that we're not there yet.