 Good morning, thanks for getting up Today, I've got a probably too much to talk about I'm going to present a two-part talk the first part Has to do the paper we published last spring in GRL and I'll be going through that relatively quickly And I'll urge you to have a read it talks about decadal mode and sea surface temperatures in the South Pacific and ocean heat content And the second part is something I'm working on right now, and I think I have some questions I'm going to show you and I'm actually interested in some feedback on the second part. I'm looking for some help So this is a figure you may have seen from I borrowed from Trenberth and Fasulo a paper where they they simply took the surface air temperature difference between The highest period in the period of the PDO before that and this is the change in Surface temperatures and particularly in the Pacific this pattern is almost identical to the PDO and it sort of highlights the importance of the PDO may be in this process of Heat sequestration in the ocean. I'm going to be talking today about our crawl records from Fiji Tonga and Raritanga That's the first part of this talk We've got three different long strontium calcium records of temperature on the time of the decadal mode on that I'll be comparing them to some records on the Pacific and also to this site in New Caledonia over here This is already been talked about by Clara Dester But one of the rationales for doing this kind of paleo work with corals is that there's a lack of them SSE observations particularly in the South Pacific this shows the number of I co ads measurements in The boxes for Fiji Tonga and Raritanga and for about 1960 there's big gaps There's actually no data during the World War one in World War two and the same also applies to the North Pacific I'll point out there's more data in North Pacific But there's still some issues with how much data there actually is to even reconstruct the PDO index So I've got a couple of figures here about what corals are and how we use them This is a very large coral from American Samoa we cored Fairly just hadn't seen this. This is a was a fish for scale over here These corals are only alive on the outer edge about an upper centimeter of very hard Skeleton and tissue we call the tissue layer below that inside. They're simply dead skeleton. It's bone white We take cores out of the middle We're very careful to plug the holes and we're done and they heal over Sometimes we can core corals that are very shallow. This one was in Tonga It's actually one of the cores we used in this reconstruction This coral had died in the 1982 83 El Nino event when sea level dropped about 30 centimeters in Tonga But we had live tissue on the sides and usually we avoid these things but actually this worked out really well We're able to take a core out of the side here And we splice it into the top of the the dead part and this is the top of that and it runs down into the 17 1790s You can see the these are density bands and extra. These are annual density bands They grow about a centimeter a centimeter half a year We these white little lines are our sample paths where we drill out every millimeter typically Of coral powder to do geochemical analysis We do not use the banding in our lab to generate the chronologies I'll talk about that in a minute because sometimes there's sub annual banding So we were lying reconstructing the annual cycle and water temperature to generate our chronologies This is the paper that came out in JRL in April. I encourage you to have a look at it There's a link on my website. It's more information about it I'll be going through this the first half of the talk right now So the main tracer I'll be talking about today is strontium calcium The strontium calcium ratio in core coral aragonite for thermodynamic reasons strontium substitutes for calcium into the aragonite lattice in the Chemistry is such that when the temperature goes up the amount of strontium goes down This that's a prediction and in many corals We find nice calibrations like this one I'm showing you at raw or tanga where that Relationship seems to hold up really well where we have a pretty tight relationship between The strontium calcium ratio and sea surface temperature. So the red line here is temperature From the I gots data set in the The black dots are our strontium calcium measurements and millimoles per mole. You can see the ratio is Decreasing upward the green data is actually a thermometer. We stuck in the water next to this coral 18 meters We want to verify that the at 18 meters depth that actually we're getting the same amplitude of the temperature cycle and we were and And then we the other thing we're doing is we tune their chronologies Let's say, you know, we tie these points typically two or that point in the two points per year to the annual cycle So that's a How we generate our chronologies, but you can even see regardless of that if we did an annual average through this Both of these we'd have a high correlation between Strontium calcium and temperature and we use that to generate a sensitivity an equation that we can then reconstruct temperatures with So we did this at all three of these sites of a show yet Fiji Tonga and Roar time I'm not going to show you all the calibrations. They're all very similar All on this is millimeter scale data and all records and this is all the data here That's in this composite up here So each one of these cores Tonga Fiji and Roar Tonga Goes back into the 1790s. There's probably over 6,000 measurements in this and you can what I've done simply in this Composite in this paper as I did annual averages of each Each some coral you counter your averages in the night, which is the individual data here in the black line is the Average of those three annual averages and around that is our standard deviation This highlights this very weak, but I think significant decadal mode It's in the temperature in this region of the South Pacific in this region that is highly correlated with the PDO in the IPO And that's what I'll be showing you in the next couple of figures So one of the problems we have is highlighted here This is a this is what this is our F. I call it the FTR composite Fiji Tonga Roar Tonga This is our coral data and with the standard deviation around these are the two different two different data products These are first temperature products We're trying to compare it to the ICO ads So this is the ICO has data for Fiji Tonga and Roar Tonga For those three different sites, three different boxes and the ERSST is processed. You can see the temperature data sets are even have some discrepancies, although the We do get better correlations between our composite and the temperature records than we do with the individual records, which I think highlights the fact that we're increasing the Signal-to-noise ratio when we make these composite in this way if we're as long as we do it carefully so I would argue that this record here which is going back now to 1791 as a Fair representation of sea surface temperature decadal modes changes in temperature in the south in this region of the South Pacific So on the top here I've got that F2 I call the FTR composite and in black back to 1791 The blue data is when you include the storage of calcium record from Christine DeLong and New Caledonia Which is much farther to the west and not quite in the same region the correlation To the PDO or IPO goes down a little bit when you add in the New Caledonia records So I haven't included the New Caledonia record in the composite, but it's a I put it here to show you that the decadal mode is Pretty widespread across the southwest specific This is the ERSST data I haven't shown the data in World War one or two because there's no ICO as data at all during those intervals and I think the data that's been infilled is I actually think it's not correct in these boxes Down here is our FTR reconstruction just showing you its comparison to the PDO and the IPO indices and This decadal variability. So this is the annual average of the dots and there's a five-year running mean I just did through the through the data. So that's a very simple processing And I think this highlights this the decadal mode is at least in the late 20th century appears to be Have a relationship to what's going on in the north pacific The evolutionary specter of that our time series shows that the recurrence interval is around 20 years in the late 20th century and it gets a little more complicated in the before that But it's somewhere between 20 and 25 years And so I think this represents the decadal mode in the south pacific And we might be able to use this as some this is a pretty long that goes back to 1790 It's the longest reconstruction of the decadal mode so far in the south pacific Um, this is another observation from this work. This is I compared our FTR reconstruction here So you upper ocean heat content in the south pacific from the levitus at all data set And so this is the entire south pacific heat content in red dots I've got the five-year running mean as a dashed line. I've got the individual yearly measurements and here The correlation is not It's 0.45 0.82 if you do a five-year running mean I think this is interesting because it suggests that When it's warmer at the surface in the south pacific on the decadal time scales It's actually the whole upper ocean is warm We're gaining heat in the upper ocean and this seems to fit with this Mariano overturning the cell mechanism or storing heat in the ocean during some decades and releasing it during others If you compare this um, which I did in the paper to The global temperature curve up just here in green and the dotted line is this the decadal extraction of that I think there is some relationship between Some in some decades where it's cooler in the south pacific and it's warmer in the global atmosphere It's not a perfect agreement in many places, but I think there's some Evidence as other people have pointed out between this the heat sequestration in the In the pacific ocean and global temperatures So that we then took our records and we're going to compare them to now some um, decadal some coral based Data from the equator from the Nino 3 4 area I'll be looking to show you from records from Jarvis and Palmyra Miana So on the top panel is the oxygen isotope measurements from three different corals I made sort of a what I call a Nino 3 4 coral composite Which is the the black line is the average And this blue line is a nine-year smooth of that and these corals are remarkably well correlated And they obviously have a very weak decadal signal oxygen isotope values and corals are a function of temperature and slownity And at this site that's there's a mixture of both variables affecting them And the only decadal shift in the in the equator that corresponds to the south pacific is the 1976 shift Coherency analysis indicates there's no there's the other decadal modes and the oxygen isotopes are knock-out here But the bottom panel is now the strontium calcium records from a couple of these corals one from Jarvis Which was released published by thompson on 2014 And palmyra and our record at ftr is the black line So what I had to do because the the Jarvis data was only released in a detrended nine-year smooth format I don't have the raw data So I actually Processed all the other records the same way So our record has now been detrended here in the black line in a in a nine-year smoothing And notice how I inverted it so when it's warmer in the south pacific it's warmer on the Equator so on this warm and substantive is colder on the equator It's opposite and notice the blue data is the highest correlation At Jarvis Jarvis is a little bit south of the equator palmyra is a little bit north And I think this suggests that um It supports the englander all idea of the trade wind fluctuations forcing decadal changes in surface temperatures on the equator And driving these meridional overturning circulation into this both the north and south pacific tricking the south pacific and we're Sequestering heat so I think our data seems to suggest that this model is has who supports this model in terms of helping to And modulate the you know the heat content of the upper ocean in the pacific So that's part one of the talk. This is the part two This is the part i'll be looking for some comments on if anyone has any We've been working on a large coral from americans Samoa and I'll be showing you some data from the upper part Now Samoa is just a little bit farther to the north. Here's the american Samoa tau These are the sites in Fiji, Tonga and Raritanga just south. This is a the sea surface temperature Here's the warm pool. This is the sea surface slinity. You can see the south pacific convergence zone And this dotted line is the sort of a slendy front running through the area This demarcates fresh water from salty water So Samoa is not very far away, but it's quite climatically quite different And that is because um, these are the nodal lines for where the Nino Inso in the pdo or ipo correlation fields run through the area. I mean nodal. I mean zero correlation line So here's the example. Here's the uh pdo pattern as we know and this is our site in Samoa and you see that this This zero correlation line runs right through Samoa for the pdo and the same is true for enso And so you would predict that um, this would not be a good place to reconstruct enso or the pdo And you might wonder why we even collected the coral there at all, but it was a very long coral And I think there's interesting information. We can get out of this And that's the highlights. I think the corals are recording what's going on So I'll be focusing on the El Nino Interannual mode for the for the for the rest of the talk So this is a paper that came out last spring by Shen and others in nature geoscience about Doing cluster analysis on the different um flavors of enso and they came out with three different clusters mark canes and they're totally more recently the heat now thinks it's like more of a continuum, but But they had three different clusters in this paper That this this is the 18 of their El Nino events were these very strong El Nino's eastern pacific types 35 percent were the warm pool El Nino's and 47 percent were the canonical type and the dot is the american Samoa The um, there's no differences between the londinas But what's interesting to me was the fact that the large El Nino events cause cooling in the american smell area Whereas the other two types don't And just farther south in Fijian Tonga. It gets cooler and saltier during every El Nino event And so what we're seeing even though it's on the zero correlation line. There's actually there's lots of variability over time you'd predict And this is another uh result that applies to this is a paper by tine others about the spcz position It was published in nature in 2012 they looked at rainfall data and they um We're looking at these these south pacific inversion. They call them zonal events When the spcz collapses onto the equator. So during normal El Nino events The uh, which is the green shift the spcz shifts to the north and east And it gets cooler and saltier in the Fiji Tonga area and sometimes at Saboa, which is down here In La Nina it shifts the other way But during these large El Nino events like 97 98 82 83 the spcz collapses onto the equator And what they call these zonal events and these are the events. I think where we can see pretty clearly in Samoa And what's going on um Here's the here's Samoa region. Here's the south equatorial current Um south this is very it's a very salty water over here in the northern jire center We just gets it vected over here into the south pacific region of Fiji Tonga and Samoa And what happens during El Nino events is the southern part of this area gets very salty and cool And it sometimes happens at Samoa And the opposite happens during La Nina events the spcz expands So our coral data is uh shown here in the bottom panel. So the calibration this is the temperature and salinity at the site the blue is uh Temperature salinity is in red and you can see there's a quite a complicated Pattern between these two parameters Our coral data is in black and now if you make the assumption that the coral data is driven only by Temperature and salinity you can actually calculate a modeled coral deli which I call pseudo coral here And the pseudo coral is in green In the so the model actually the data fits the model pretty well the correlations are about 0.6 Given some of the chronology uncertainties we have on the annual basis. That's pretty good And I think this tells us the coral is actually recording the deli 18 of the coral is recording this complicated mixture of Temperature and salinity variability at this site and it receives you can see the big El Nino events a 97 98 82 83 Gets cooler and saltier So on the top here, I've removed I've extracted the the inter-annual band and deli 18 I've detrended the data and compared it to the Nino 3 4 index In that the arrows highlight the um the very large El Nino events where it gets cooler and saltier at Samoa and you see they're concentrated in the late 20th century And if you do a running correlation through the uh this with a 25 month um window This um interesting pattern comes out and you so that's on the zero correlation line You expect that it would it would be shifting around over time But it appears to be um shift from one mode to another in the mid 1920s and right around 1927 actually It's quite abrupt shift where the so in this part of the record we it's saltier on average and cooler in Samoa during El Nino events and before that it's warmer and fresher And the 1920s and it may should be shifting back again late early a farther back in the record And this if we compare our coral record now, which is in red which with the trend in it to the The coral composite record I showed you before We get the same exact pattern the in 1920s the mid 1920s something happened with the The correlation or the phasing of El Nino events or the To the to the tropics in the south pacific And so what I think is going on is it's getting warmer and fresher At Samoa before 1920s in Samoa during El Nino events and the opposite's going on after that This they qualify as a climatic ship. I'm wondering based on a talk on monday This is not true with Fiji just to the south. This is a record from Emily Dacier's composite paper in 2014. There's no Phase shift in the 1920s to the Nino 34 temperatures. So this is only going on at Samoa Which is to the north this corresponds to a as we've talked about a sort of switch in the AMO in the Atlantic in the 19 mid 1920s and it also corresponds to Or I'm not sure this is related, but there's an interesting The largest floods in Mississippi river ever were in 1927 when over in a 27,000 square miles of the Flood plane was up in 30 feet of water and it was the biggest flood ever and I'm not sure this is related But I wonder what's going on in the mid 1920s and what I think What I think I was going on is that people think the ITCZ has shifted north in the 1920s in the Atlantic And I think the SPCZ did the same thing. I think it was farther to the south in west Before 1927 and it shifted to the north rather abruptly and that would explain I think the pattern we see of El Nino variability at this site So to wrap up here I think the Decadal mode in the south pacific leads back to for over the 200 years is around 20 25 years in temperatures I think the supports are what I've showed you supports this Shallow returning cell mechanism for storing heat in the ocean I think our Samoa data suggests that there's some kind of abrupt change going on in the 1920s And I'm going to look more into this get some feedback from people here And I wonder if there's a coordinated ITCZ change in both basins at this time And I want to point out that I I think some of this work highlights the need to really closely look at these paleo sites Related to small scale variations in the ocean particularly at places where there's gradients Um because I think the small scale small scale variations of the oceans do matter and we need to keep to keep uh Keep track of them and we do our interpretations. So I'll stop there. Thank you