 Okay, so first I want to thank the organizers for giving me the opportunity to talk about this research today and also thank Allie for providing such a good overview of the West African monsoon, which we'll be continuing to talk about in my talk. This is work that was published last year as part of my PhD dissertation at Columbia working with both Allie and Yocanin. And as you can see from the title, it details 20th century sea surface temperature driven decadal variability in the Sahel using a moisture budget framework or analysis. So I'll just briefly gloss over this because a lot of it was already mentioned in the prior talk, but when we think about the Sahel region of Africa or the West African monsoon, really where we're focusing on is the some of the land area in the countries outlined by the black box there on the top. And this is a region of the world which is a semi-arid grassland that really exhibits a canonical monsoon climate. So what I've shown in the bottom two panels are just the January, February, March and the July, August, September rainfall and wind climatology. The winds are for the 925 millibar level. So we see this really clear shifting of the winds from the winter season to the summer season, which of course brings this moisture to the area and helps to maintain that rainfall that we see in July, August, September over the region. And in terms of Sahelian decadal variability, as we as has been mentioned many times this week, it's kind of this poster child for decadal variability over land. Here I'm showing the precipitation anomalies and various data sets, most notably the University of East Anglia precipitation data set, which is station derived estimates that are then interpolated to a latitude-longitude grid. And it's pretty clear to see over the 20th century this shifting in the precipitation from the relatively wet 1950s and 60s to this really intense dry period in the 70s and 80s. So superimposed on top of the gray bars which is just the raw precipitation estimates is this decadal signal which I obtained from just applying a 10-year running mean to the data. And really this shift was unseen in other parts of the world in both its magnitude and spatial extent. And as Ali was pointing out, we've now robustly shown that it's really driven to a large part by oceanic forcing. However, in terms of different AMIP-like studies, various authors have pointed to the roles of different ocean basins. And so in this work what we decided to do was to revisit the sea surface temperature forcing in the Sahel and not just look at the precipitation field but actually do a full moisture budget decomposition to try to get a better understanding of some of the mechanisms that are at play. And within this framework we're able to study the moisture budget terms and really link their changes to oceanic forcing because we're only looking at this with the CAM4 ensemble mean variable. So we had a 16-member ensemble mean using historic sea surface temperatures for the entire 20th century 1901 all the way to 2008. We take the average of that and we look at the different terms both in their climatology and in their decadal variability. So I can get the pointer to work. Okay, what I have circled are different terms of the moisture budget equation. This is the only equation I have in my slideshow. So we have net surface moisture which is precipitation minus evaporation on the left-hand side of the equation is equal to the moisture convergence by the mean flow, the first two terms following the first summation and then moisture convergence by the transient flow. So here we're really interpreting the mean flow convergence as being due to those monsoon-west release whereas moisture convergence by the transient are can be interpreted as sub-seasonal variations and just based on vector algebra you can you can break down the moisture convergence by the mean flow into a portion due to mass convergence. So that's the first term on the left-hand side of that first summation and then the specific humidity advection. So we want to do this accounting and again look at the climatologic behavior and also the decadal variability and try to see how that's consistent with various ocean forcing. So the first part of the analysis was to validate the AMIP CAM4 model output by doing a maximum covariance analysis. This shows us the covariations between the global sea surface temperature field and Sahel precipitation shown for the observations on the left and for the model on the right and what we can see is this really strong agreement between the two different types of data where we have drying in the Sahel that's really forced by the strongest sea surface temperature anomalies which are again that meridional gradient across the Atlantic when you have a relatively cool North Atlantic compared to a relatively warm South Atlantic you tend to see drying in this area and also a warm Indian Ocean. So after doing that in the paper and also validating the model with some other techniques we go ahead and as I was saying before compute these moisture budget terms both in the climatology this is for July August September of course when it's the monsoon season and later for the decadal variability and here there's many different panels for the different terms but I'm just going to highlight a few that really stuck out to us. Green colors denote convergence or precipitation of course while brown is divergence and what we can focus on is this is the moisture convergence by the mean flow term with which if we compare to the above two panels one of which is the net surface moisture and the total moisture flux convergence we see that it's really the moisture convergence by the mean flow that's dominating both the magnitude and the spatial patterns of the rain band in Africa whereas the transients which can be seen to the right of that panel are really only working to weekly wet the Sahel in a climatologic sense. And next if we move to the bottom panel to consider the mass convergence it's really the mass convergence locally that's giving the structure to that mean convergence field and therefore also the total and the net moisture at the surface whereas the specific humidity advection term in an overall climatologic sense is actually drying over the Sahel. Now we can keep those plots in mind which are also actually represented by the contours in these next panels and we then look at the moisture budget anomalies for July August September which are regressed on that decadal index that we obtained from the maximum covariance analysis technique. So each of these panels is kind of can be interpreted as the change in these terms that's attributed to that sea surface temperature forcing that was the major mode of variability so again the meridional gradient in the Atlantic and a warm Indian Ocean. The same eight panels for the different terms in the moisture budget equation and again I'll just highlight a few that stick out to us they're not surprisingly some of the same terms from the climatology but the moisture convergence by the mean flow here seems to show a bit more east-west dichotomy particularly you see some convergence of moisture to the east of Lake Chad that you don't see when we look at the west and if we look down again at what the mass convergence is doing if we looked at that in isolation we might assume that okay these sea surface temperature patterns are actually forcing a lot of convergence of mass particularly in East Africa that if we thought about that purely alone we would say we would we might think that it should become wetter in that area but of course that we know that these terms are operating together and really the specific humidity advection in this case so driven by those changes in the ocean seems to have a larger impact on what's happening in the eastern most part of the Sahel more so than the west and really combines with the other terms such that we do see drying overall but it's particularly pronounced in West Africa due to changes in circulation versus in East Africa perhaps because of this specific humidity advection term dominating so moving on what we tried to do next was to see whether some of these convergence patterns were consistent with the monsoon dynamics that we obtained also from the model ensemble and so what we've done is taken a longitudinal average from 10 degrees east to 10 west that's shown in the top of the figure so where the drying signal is the strongest and we've plotted on the x-axis here latitude and on the vertical the pressure level in at least the middle panel for the windfields the bottom is the sea level pressure and the top is just a precipitation histogram so in terms of the climatology this tells us something about what the dynamics look like where you have strong precipitation strong moisture convergence and the difference between a dry and a wet decade tells us how these terms are operating in an anomalous sense and again forced by the ocean given the experimental design so just very quickly in terms of the climatology we see this canonical monsoon peak where you have the highest rainfall rates occurring between approximately 10 and 12 and then extending to 18 to 20 degrees north which is the northward extent of the Sahelian rain band this is fed by low-level southerlies from the Atlantic Ocean feeding that area with moisture and we also see the a maximum region of ascent just over that same peak in the rainfall in terms of latitude this coincides with a surface low pressure center that happens between about 20 to 24 degrees north and as known in the literature as the Sahelian heat low and has been shown to be really important in terms of setting up this monsoon and really maintaining it and if we take those considerations and then think about the differences that are displayed in the dry compared to the wet periods what we see is not only that the rainfall reductions are maximized in that main rain band and not at the margins but we do see consistent dynamical behavior such that there's a reduction in the low-level monsoon southerlies that are feeding the monsoon but there are also increases in the wind field from the north so this intrusion of drier from the Sahara does seem to increase during the dry period compared to the wet which would be consistent with what the bottom panel is showing which is this filling in of the Saharan heat lower weakening of the low so then we decided to look at some of the main moisture budget terms again to see if these dynamics were consistent and we saw some of these things being played out with the moisture budget terms so again these are the same panels the climatologic behavior and that on the left and the behavior in dry minus what periods on the right for a latitude pressure cross section and what we see in terms of the climatology is that you have low-level surface conversions that does occur much further north than the maximum latitudinal extent of the rain band so again that's about 18 to 20 north but this is more than compensated for by upper-level divergence which is in some part attributed to the mass convergence that's the middle panels but really it's the specific humidity advection at that northward the northward extent of the monsoon so we're attributing that to the flow coming in from the Sahara desert that's really setting the northward migration and drying out the land at that at that area and a lot of the patterns are consistent in the dry compared to the wet period as well where you see the bottom right panel the specific humidity advection really plays its strongest role again just northwards of the monsoon rain belt and helps to set the migration of the monsoon whereas the mass convergence itself is what's dominating where it's raining so again between approximately 10 and 18 or 20 degrees north so in conclusion we were able to show in this paper that decadal scale variability is realistically represented by this AMIP run using CAM4 and specifically it verifies like other studies have shown the dominating role of a warm Indian ocean and a meridional gradient across the Atlantic and being one of the largest drivers of drought in this region during the 20th century increased moisture divergence by the mean flow is what accounts for most of the drought whereas you have some weak drying by transients during the 20th century and finally in terms of the main rain belt you can think of the changes in mass convergence as being the dominant source of that the variations whereas advection of specific humidity we can interpret as potentially setting the northward migration of the rain belt so thank you for your attention I'll take questions and I encourage you there's much more detail in the paper if you'd like to read more