The MWP and LIA in Kenya and South Africa

In contrast to the claims of the Mann et al "Hockey Stick", practically all proxy records around the world show large scale multi-decadal and century-scale climatic variation.

As another example, here are some of the results from a study by Tyson, Holmgren et al (2000,2001) of Lake Naivasha in Kenya and a well documented stalagmite in a cave at Makapansgat, South Africa. The principle aim was to show a "teleconnection" between the two sites which are far apart, and show almost complete climatic anti-phase between each other.

The most interesting result is the imprint of the Medieval Warm Period and Little Ice Age in Africa. During the MWP, Lake Naivasha actually dried up for more than two centuries, and during the Little Ice Age, Lake Naivasha was much deeper, with the modern depth of the lake somewhere between these two extremes.

Here’s one of the diagrams to show the teleconnection between the two sites: Lake Naivasha and Makapansgat Valley proxy reconstructions

The study reports:

Depleted 18O values in the Makapansgat stalagmite record are associated with drier conditions and increased occurrence of deep, intense thunderstorms and hail; enrichment occurs with wetter conditions and persistent warm rainfall from middle-level stratiform clouds (Holmgren et al., 1999). Similarly in the ´13C record, humate-induced changes in colour and growth rate correspond to changes in vegetation cover and temperature that indicate, in general, that drier periods were cooler, whereas wetter times were warmer (Holmgren et al., 1999). In the past at Naivasha, during centuries-long droughts, when rainfall, runoff and lake levels were at their lowest, conditions were warmer and wetter at Makapansgat. The climate gradient towards more mesic conditions in subtropical southern Africa reversed from time to time and was most pronounced during the medieval warm period (900 1300 AD) and during the Little Ice Age (1300 1800 AD). A clear correspondence between moist conditions and periods of solar inactivity (theWolf, Sporer and Maunder Minima) has been observed in the Lake Naivasha region (Verschuren et al., 2000), with highest lake levels in the last 1100 years having been recorded during the Maunder Minimum. In the Makapansgat record exactly the opposite occurs, with coolest and driest conditions occurring during the Maunder Minimum (Tyson et al., 2000).

Once again, the natural variation of climate is shown in proxy records and closely follows variations in the solar cycle over the past 1000 years and more.

Update, checking a tourist site on Lake Naivasha, I found this:

Lake Naivasha is a beautiful freshwater lake, fringed by thick papyrus. The lake is almost 13kms across, but its waters are shallow with an average depth of five metres. Lake area varies greatly according to rainfall, with an average range between 114 and 991 sq kms. At the beginning of the 20th Century, Naivasha completely dried up and effectively disappeared. The resulting open land was farmed, until heavy rains a few years later caused the lake to return to existence, swallowing up the newly established estates.

So Lake Naivasha does have a tendency to dry up for short periods of time. What does stand out in the reconstruction is that during warm periods Naivasha dries up and during cold periods, the lake gets deeper as the local climate gets wetter. Naivasha does not have an outlet for its water, making the depth of the lake a key function of the wetness/dryness of the Kenyan climate.


  1. Joanne Ballard
    Posted Mar 1, 2005 at 4:18 PM | Permalink

    I am just wondering about this correlation. It’s very interesting.
    What I’m wondering is, are these accurate interpretations? Do we have data
    of other lakes in recent times drying up, and caves in the same vicinity (perhaps Western
    U.S.A.) so that we can confirm the proxy data on stalagmites and lakebed sediments with recorded data? I have not taken classes in climatology, so this is just my idea of checking validity. Perhaps climatologists are pretty sure of the correlation because they understand weather patterns, air, moisture and lake evaporation better than me. Is a dry climate the only parameter for seeing depleted 18O in a stalactite or lake sediment core?

    John writes: To be honest, I don’t know. My impression is that depleted 18O is an indication of lower solar irradiance, and hence lower global temperatures.

    What I am trying to do (and readers, feel free to contribute) is produce evidence from all over the globe of climatic change over the past 1000 or more years, and compare this with the picture shown by the Mann Hockey Stick. Although Mann (and mouthpiece David Appell) say that the Medieval Warm Period and the Little Ice Age were regional and not global, I cannot see that his reconstruction can realistically claim this if the same imprint is shown from all parts of the globe. More importantly, the Hockey Stick is an agglomeration of regional climates – it is not in the same league of global coverage that satellites are today.

  2. Michael Ballantine
    Posted Mar 1, 2005 at 10:28 PM | Permalink

    These, I believe are useful climate indicators. And yes, we should look for similar records. Will other lakes follow the same pattern? Maybe and maybe the reverse or maybe not at all. It all depends on how the global temperature affects that region.

    And isn’t it true that ALL proxie indicators record how their region responds to the global temperature?

    And I will ask here something mentioned in another forum. We see lots of excitement over a global temperature rise of about one degree with arguments over tenths of a degree. If we deal strictly with the instrumental record for the moment, does that mean that all of these temperatures are recorded to the nearest 0.1 degree? Are all of these instruments calibrated and correlated to a NIST standard every year? If the temperatures are only recorded to the nearest degree then the real average temperature could be +/- 0.49 degrees. That is a range of 0.98 degrees with all values having equal probablitiy. It is not a bell curve or any other center weighted curve. This questions intentionally ignores the placement of the sensors and whether someone has built anything near them or cut any nearby trees over the recorded period.

    John writes: now you come to the biggest problems of climate science: the effects of urbanization, the failure to have a global calibration standard, the lack of quality in third-world temperature stations. It could be argued that + or – 0.1C is a lot less than the standard error of the thermometers used.

    Dr John Christy tried to get the temperature record for just North Central Alabama. It’s more complicated than you’d think.

  3. Louis Hissink
    Posted Mar 2, 2005 at 5:00 AM | Permalink

    Michael has hit the proverbial nail on the head – apart from the more serious problem of assuming that temperature measurements are actually some physical countable quantity (intensive vs extensive variables, a concept I discovered is certainly not at all well understood).

    The errors are accentuated when temperature anomalies are computed too, though how theoretically random errors of reading lead to such convincing graphs depicting historical temperatures makes we wonder whether Steve has more fertile ground to plow.

  4. Joanne Ballard
    Posted Mar 2, 2005 at 6:06 AM | Permalink

    Michael: “And isn’t it true that ALL proxie indicators record how their region responds to the global temperature?”
    (Sounds like Perry Mason 😉 ) A one size fits all given? It’s a nice neat concept. But IS precipitation that uniform isotopically regarding temperature? We see pretty dramatic spikes in ice cores, ocean sediments, lake sediments, stalagmites, and even mammoth tusks. Can there be other factors? I’m just curious and want to test these data for myself in some way.

    Yes John, I can understand that. André, in ukww, has done some pretty extensive investigations on urban heat islands vs. rural stations, and found some interesting things with that study.

  5. Frank H. Scammell
    Posted Mar 2, 2005 at 9:02 AM | Permalink

    It is my understanding that the “instumental record” (i.e., GISS temps.) is pretty heavily massaged from the “true” thermometer readings (to accommodate “urban heat islands, etc.) and shows an upward drift that is not evident in the satellite (and balloon) records (either troposphere or stratosphere). Thus, the upward drift cannot be confirmed. This drift (viewed by some as “exponential”!) may not be real, but it is included (in some form or other) in all of the “respected” models.

  6. Joanne Ballard
    Posted Mar 2, 2005 at 9:42 AM | Permalink

    I have seen a graph of ice core data vs. snow height, showing a positive correlation between dD and snow height. It seems that if it snows more, the heavy hydrogen levels go up. Surely someone has looked at this before. Could it really be that simple?

    John writes: I’d like to see that study. Please send it to climateaudit AT

  7. Andre Bijkerk
    Posted Mar 3, 2005 at 1:26 AM | Permalink

    John, In the studies (Science mag) about the ice core cores about the Bolling-Allerod-Younger Dryas-Pre-boreal sequences the dramatic change of precipitation is noted but nothing more. I never have seen a study noting the extreme correlation between accumulation rate and dD isotope rate thoughout the full 420,000 years of the Vostok Ice core (Petit et al 1999, 2001). Yes it seems to be that simple. Try this thread:

  8. Louis Hissink
    Posted Mar 3, 2005 at 5:30 AM | Permalink


    Interesting link to the thread – I have not yet bothered with downloading the data, but from experience are these different measurements derived from the same source – vis Vostok Ice Core?. The similarity of responses from the vearious measurements suggests to me they are quantifying sampling artefacts rather than intrinsic factors. Gut feeling, that is all.

    We get the same in geochemical exploration – case history – decades ago we had a geochemical team collecting soil samples, (6 month job) and initial data came from the lab. I started plotting the data up and noticed that values were pretty consistent, too consistent. Correlating with known geology showed that at least the assays (Cu, Ni, Cr, Co) shoudl have matched our mapping of the rock types, ultramafics, mafics, felsics etc.

    Lab manager telephoned a few days later expressing doubts about the analyses – which prompted an unannounced field checl to discover the crew filling the sample packets from a 44 gallon drum.

    As Steve would confirm, we mining types have acute radars for “unusual” data.

    I would urge caution in interpreting the data in this thread as a result. It may not be THAT easy.

  9. Peter Hearnden
    Posted Mar 3, 2005 at 5:59 AM | Permalink

    Frank H. Scammell wrote

    “It is my understanding that the “instumental record” (i.e., GISS temps.) is pretty heavily massaged from the “true” thermometer readings (to accommodate “urban heat islands, etc.)”

    Indeed, the planet is warmer than the figure show…I wonder what the reaction would be if I used the word ‘massaged’ re what Steve is doing? Put up or shut up I think.

    “and shows an upward drift that is not evident in the satellite (and balloon) records (either troposphere or stratosphere). ”

    Try for another interpretation of the raw sat data.

    “Thus, the upward drift cannot be confirmed. This drift (viewed by some as “exponential”!) may not be real, but it is included (in some form or other) in all of the “respected” models”

    But, the upward drift is confimed by RSS. Which is why I bet you’ll dismiss them. Am I right?

  10. Joanne Ballard
    Posted Mar 3, 2005 at 10:22 AM | Permalink

    Hi Louis, Yes of course it is right to be skeptical on this. The ice cores are drilled and stored. Then someone has to collect and enter that data into a database. How did they decide what snow height would be? The width of the bands of dark and light I would think. As I understand it (haven’t seen a real ice core yet), there are dark and light varves representing winter and summer. Those layers get compressed over time from snow to firn (old snow) to ice (maybe a l00 years is needed for that)as more snow accumulates at the surface.

    Steve: Joanne, on accumulation, look at the post from Hans Erren on Quelccaya accumulation.

  11. andre bijkerk
    Posted Mar 3, 2005 at 11:54 AM | Permalink


    The point is that there is no interpretation. Just plain basic math with the results in a graph. That’s it, no suppositions whatsoever. Just distance of sampes divided by the delta age, corrolated to the heavy isotope isotope rate in the same sample. Nothing more, nothing less. You would not expect a correlation over 90%, but yet, its there. Once the evidence is correlated to something you cannot use the same for something else, like temperatures (directly) for instance. Indirectly okay, considering the relation between moisture and temperature. Consequently the big spikes in the ice cores are not about 10-12 degrees temperature change.

    The ice age was totally different

  12. Jeff Norman
    Posted Mar 3, 2005 at 2:13 PM | Permalink

    Peter Hearnden,

    Your point regarding the RSS data is valid. Unfortunately it still doesn’t corroborate the modelled temperature increases presented in the IPCC story lines.

    The CRU global temperature summary indicates an increase in surface temperatures of ~0.45 K (1979 to 2004). This is reported to include a correction for the urban heat island impacted weather stations included in the proprietary data mix at the CRU.

    The UAH global temperature summary indicates an increase in tropospheric temperatures of ~0.11 K (1979 to 2004).

    The allegedly improved RSS global temperature summary indicates an increase in tropospheric temperatures of ~0.32 K (1979 to 2004).

    Assuming the RSS interpretation of the MSU data is correct, you have to conclude either:
    1. The Earth’s surface is not warming as fast as the CRU concludes; or
    2. The surface is in fact warming faster than the troposphere and therefore the anthropogenic global warming hypothesis is incorrect.

    What is interesting about both the UAH and RSS MSU datasets is that they record a global tropospheric temperature increase in 1997/1998 and a subsequent global tropospheric temperature decrease in 1998/1999 that is larger (up or down) than the total temperature increase recorded in the same 25 year period at the CRU.

    The temperature increased due to the 1997 el Nino. Huge amounts of heat energy were transferred to the atmosphere. Where did the heat subsequently go? What caused the temperature decrease?

    If the atmosphere can shed so much heat so easily (<18 months), what does this say about the greenhouse effect?


  13. Joanne Ballard
    Posted Mar 4, 2005 at 12:35 AM | Permalink

    Steve, can you point me to Hans post on the Quelccaya ice? Thanks, Joanne

    John writes: Check out this thread which links to Hans’ work.

  14. Hans Erren
    Posted Mar 4, 2005 at 3:42 AM | Permalink

    Joanne, It’s here

  15. Louis Hissink
    Posted Mar 4, 2005 at 7:00 AM | Permalink


    That is how it was calculated?.

    Visual classification of ice core classes these layers into events, more probably related to depositional ones, rather than chronological.

    Tends to confirm my initial interpretation that the measurements reflect sampling rather anything else.


  16. Hans Erren
    Posted Mar 4, 2005 at 8:39 AM | Permalink


    IMHO using lithostatigraphy in a seasonally determined sedimentation regime is justified. Please show us the caveats of this assumption.

  17. Louis Hissink
    Posted Mar 5, 2005 at 10:40 PM | Permalink


    is not the layering principally one of indiuvidual snow storms, of which many might occur in one year? I recall reading elsewhere (references are 3000 km away in Sydney which I will get my hands on over Easter) relating to this very issue. I have to leave this particular topic until I refresh my memory on the particulars – but I do recall the debate was pretty acrimonious, neither side giving in. In the meantime I might have another obscure reference to it somewhere in Perth, which I’ll search.

  18. TCO
    Posted Sep 17, 2005 at 11:48 PM | Permalink

    I’m not sure how useful it is to show this kind of study. It begs the question of “is this proxy representative of a majority or minority”? Also, its not being exposed to the same full court press of looking for confounding factors as the bristlecones, etc. are. There’s also a comment by JohnA about looking at studies all over the world to get some idea if MWP occured or not (in essence an overarching study). However, that does not seem to have been done. Instead snippets that support one position are brought forward.

    Then there is the blablabla about heat islands jacking the thread…

  19. JillB
    Posted Aug 7, 2006 at 6:10 AM | Permalink

    A site index has turned up on the left hand site of the page,I haven’t seen it before in the previous 6 or so weeks when i have dipped in. It obscures most of the text even though i have put the font to smallest. How do i get rid of this index so i can read the blogging.

  20. fFreddy
    Posted Aug 7, 2006 at 6:27 AM | Permalink

    JillB, this arises when people put URLs early in their posts so that they appear in the recent posts section at the top right. The problem only applies to MS Internet Explorer users, because of the non-standard way MS tries to do things. The best solution is to use Firefox, which is much better for a whole host of reasons.

  21. KevinUK
    Posted Aug 7, 2006 at 8:36 AM | Permalink

    #19 and 20

    John A, Is this side effect of long URLs at the start of a post down to WordPress or is the index rendering code yours?

    I presume its the former, if the latter then I’d be happy to help solve the prolem.


  22. bender
    Posted Aug 7, 2006 at 8:39 AM | Permalink

    Please do!! Thanks KevinUK!

  23. Henrik
    Posted Oct 4, 2006 at 6:11 AM | Permalink

    Antartic is not the place to start comparing temperatures between current warm period and medieval warm period. As Nasa satellites tell us, Antartic has had cooling trend for a long time. I’m not surprised if there is correlation between cooling Antartic and warming NH especially Europe and N.America.

    It could be possible and “global” cooling or warming happens or have happened in past mostly only in NH. There is quite strong “concensus” with european paleoclimatology that top MWP temperatures were higher than current peak temperatures. Treelines (oak, pine, spruce ) in Scandinavia, Alps and icecores are the best proxies. It tells MWP (about 1000-1200 AD) temperatures were about 1 degree higher than late 1900’s. I’m not impressed of any isotope proxies, cause these are problems with timings and so on. They have some risks.

One Trackback

  1. By The Trouet Ocean Proxies « Climate Audit on Sep 11, 2010 at 10:34 AM

    […] sediment series, mentioned in passing in connection with Mann et al 2008 here , also here. The data set was also used in the discussion of a warm Warm Pool MWP by Newton et al 2006 (See CA […]

%d bloggers like this: