Quantitative predictions of past climates are key to the reconstruction of palaeoenvironments and are essential for climate model validation. Iron oxide nanoparticles form as a by-product of the pedogenic processes that occur during soil formation. These particles reflect the environmental conditions under which soil formation took place and can be quantified using standard rock magnetic measurements. Numerous quantitative predictions of past climates have been based on soil magnetic mineral assemblages. For example, detailed time series of Quaternary mean annual precipitation and palaeoprecipitation gradients across wide geographic regions have been predicted from the rock magnetic properties of Chinese loess and palaeosols (Maher and Thompson, 1995). However, little attention has been given to estimating the uncertainties associated with such empirical rock magnetic climofunctions and the predictions they make (Figure 1).
We have performed a retrospective error analysis on an ensemble of published climofunctions that were previously employed to predict past rainfall levels from a variety of soils on different continents. Based on our analysis we can now assign uncertainties to a number of the palaeoprecipitation predictions published in the literature. We find that existing climofunctions have associated uncertainties that are so large their resulting predictions are effectively invalid. Thus, current palaeoprecipitation reconstructions must be treated with extreme caution.
Our findings suggest that future climofunctions should incorporate more geological considerations. By taking into account both empirical and theoretical models of soil development, smaller uncertainties can be achieved by the construction of physically-constrained climofunctions.
References
Balsam, W. L., Ellwood, B. B., Ji, J., Williams, E. R., Long, X., El Hassani, A., 2011. Magnetic susceptibility as a proxy for rainfall: worldwide data from tropical and temperate climate. Quaternary Science Reviews, 30, 2732-2744.
Maher, B. A., Thompson, R., 1995. Paleorainfall reconstructions from pedogenic magnetic susceptibility variations in Chinese loess and paleosols. Quaternary Research, 44, 383-391.