EGU General Assembly 2020
© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Northern Hemisphere temperature to precipitation relationships during the last Glacial from pollen records and climate simulations

Anna Sommani, Nils Weitzel, and Kira Rehfeld
Anna Sommani et al.
  • Heidelberg University, Institute of Environmental Physics , Physics Department, Germany (

The hydrological response to radiative forcing is less understood than the thermal one: many climate models have difficulties in simulating seasonal rainfall and its variability. Indeed, future precipitation projections are much more uncertain than those of temperature. However, confident projections of precipitation are of crucial importance, particularly for highly populated regions where agriculture strongly relies on seasonal rainfall, such as South and Central Asia.

Instrumental data from Eurasia show a negative correlation between temperature and precipitation on short timescales (10-3 to 100 years). However, on longer timescales (101 to 103 years), proxy data covering the Holocene show a positive correlation between temperature and precipitation. Climate models in contrast simulate a negative correlation on all timescales. To extend previous estimates to longer time scales, we focus on the last Glacial period, characterized by colder temperature than the Holocene as well as pronounced millennial-scale climate fluctuations in the Northern Hemisphere.

We reconstruct temperature and precipitation from four high resolution pollen records at mid-latitudes in the Northern Hemisphere. The estimates are compared with climate simulations. The chosen proxy sites cover the East and West coasts of both the Eurasian and North American continent. We employ four different statistical reconstruction methods to assess validity and biases of each method. The differences between reconstructed and simulated temperature-precipitation relationships as well as the zonal structure of orbital- and millennial-scale variations are examined. In particular, we explore the thermodynamic and dynamic contributions to the inferred relationships between temperature and precipitation.

How to cite: Sommani, A., Weitzel, N., and Rehfeld, K.: Northern Hemisphere temperature to precipitation relationships during the last Glacial from pollen records and climate simulations, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-7293,, 2020


Display file

Comments on the display

AC: Author Comment | CC: Community Comment | Report abuse

displays version 3 – uploaded on 04 May 2020
Citation correction.
  • CC1: Comment on EGU2020-7293, Sandra Gomes, 05 May 2020

    s welDear Anna, 

    Thank you for sharing this well organised and pertinent information. I work with marine pollen records, high-resolution Last glacial. I like to follow the developments/progress to reconstruct temperature and precipitation. I would like to ask two questions, and sorry in advance for my ignorance:

    a) what is the main advantage of using WAPLS?

    b) did you take into account the effect of CO2, which was limiting factor for plants development during the last glacial, and can bias the parameter reconstructions? as well as the obtained correlations?


    Thank you



    • AC1: Reply to CC1, Anna Sommani, 08 May 2020

      Dear Sandra,
thank you for you questions, both of them point to very important aspects.


a) In our preliminary climate reconstructions we use WAPLS because it is a well established and popular method for reconstructing environmental variables from species assemblages.
 However, we are not only focusing on the WAPLS method, but also testing and quantitatively evaluating and comparing other methods than the WAPLS (WA, MAT, MLRS and Bayesian models). In general, the main advantages of WAPLS compared to, MAT, is that WAPLS is able to perform well even in ‘non-analogue’ situations, and, compared to WA, WAPLS overcomes the typical ‘edge effects’ of WA reconstructions. 
If you are further interested in the advantages and disadvantages of climate reconstruction methods, we would strongly recommend you this article, “Strengths and Weaknesses of Quantitative Climate Reconstructions Based on Late-Quaternary Biological Proxies”, Birks, H John B Heiri, Oliver,Seppä, Heikki, and Bjune, Anne E.

      b) In the preliminary analyses presented here we do not account for influences from the changed CO2 concentrations, because the standard reconstruction methods do not provide ways to correct for it. However, we are aware of the potential biases from it and we will try to account for it in later stages of the project.


displays version 2 – uploaded on 04 May 2020, no comments
Corrected one typographical error.
displays version 1 – uploaded on 04 May 2020, no comments