ENSO continuum and its impacts on worldwide precipitation: Observation vs. CMIP5/6
- 1Coventry University, Centre for Agroecology, Water and Resilience (CAWR), Coventry, UK (bastien.dieppois@gmail.com)
- 2Department of Oceanography, MARE institute, University of Cape Town, Cape Town, South Africa
- 3Department of Meteorology, National Centre for Atmospheric Science (NCAS), University of Reading, Reading, UK
- 4CRC/Biogéosciences, CNRS/Université de Bourgogne Franche-Comté, Dijon, France
- 5Science Partners, Paris, France
- 6Department of Geography, Kong Kong Baptist University, Hong Kong, China
It is now widely recognized that El Nino-Southern Oscillation (ENSO) occurs in more than one form, e.g. eastern and central Pacific ENSO. Given that these various ENSO flavours may contribute to climate variability and trends in different ways, this study presents a framework that treats ENSO as a continuum to examine its impact on precipitation, and to evaluate the performance of the last two generations of global climate models (GCMs): CMIP5 and CMIP6.
Uncertainties in the location and intensity of observed El Nino and La Nina events are assessed in various observational and satellite-derived products (ERSSTv5, COBESSTv2, HadSST1 and OISSTv2). The probability distributions of El Nino and La Nina event locations, and intensities, slightly differ from one observational data set to another. For instance, La Nina events are more intense and more likely to occur in the central Pacific using COBESSTv2. All these products also depict consistent decadal variations in the location and intensity of ENSO events: i) central Pacific ENSO events were more likely in the 1940/50s and from the 1980s; ii) eastern Pacific ENSO events were more likely in the 1910/20s and 1960/70s; iii) La Nina events have become more intense during the 20th and early 21st centuries.
These fluctuations in ENSO location and intensity are found to impact precipitation consistently across diverse global precipitation products (CRUv4.03, GPCCv8 and UDELv5.01). Over southern Africa, for instance, more intense eastern (central) Pacific El Nino events are found to favour drought conditions over northern (southern) regions during austral summer. By contrast, over the same regions, more intense La Nina events favours wet conditions, while the location of these events has little effect on precipitation. Over West Africa, ENSO locations favour a zonal (E-W) rainfall gradient in precipitation during boreal summer, while changes in ENSO intensity modulate the strength of the meridional (N-S) rainfall gradient.
Using both historical and pi-Control runs, we demonstrate that most CMIP5 and CMIP6 models favour either eastern or central Pacific ENSO events, but very few models are able to capture the full observed ENSO continuum. Regarding ENSO impacts on worldwide precipitation, contrasted results appear in most models.
How to cite: Dieppois, B., Eden, J., Monerie, P.-A., Pohl, B., Crétat, J., and Chun, K. P.: ENSO continuum and its impacts on worldwide precipitation: Observation vs. CMIP5/6 , EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-9606, https://doi.org/10.5194/egusphere-egu2020-9606, 2020.
Comments on the display
AC: Author Comment | CC: Community Comment | Report abuse
Bastien,
As I said during the session, I'm surprised by the very large observational uncertainties in ENSO location. Do you use the entire 1860-2018 period? Did you try to do the same analysis with a shorter period (e.g. starting in 1950), and does it change the conclusion about the model / observation agreement? I suppose that the SST measurements in the tropical Pacific were very rare in the late 19th century and early 20th, right?
Hi julien, yes sorry I did not had much time to answer.
Yes I used the data between 1860 and 2018. But, I did look at different time periods. On slide 3 the bootom panel is the agreement to show high-probability of events between the data set, and over time.
The late 19th century wasn't that bad, I guess all data sets use the same data from interpolation and extrapolation at that time (even though they are fewer). They are more conflict during World War 2. On this panel, you can also see that the large difference between OISST and the other data sets is mostly due to the time period. It sounds like central Pacific ENSO are much probable over the last 30 years (hence the discrepancies in the upper panel). However, I don't know why eastern Pacific ENSO are less probable in ERSST than in COBESST and HadSST. I guess this might be due to the use of satelitte data in the former two. in some ways it shoudl be more accurate using satelitte data, but in another way it's introducing large inhomogeneity in the "assimilation" process, and a cold bias as highlighted in Reynolds et al. (2002).