Simultaneous growth releases and reductions among Populus alba as an indicator for floods in dry mountains (Morocco)

Abstract

We studied the growth reaction of silver poplar trees (Populus alba) to a large flood in November 2014 in the semi-arid High Atlas Mountains, Morocco. The flood resulted in half of the studied trees developing wider tree rings in 2015 and the other half developing narrower rings in 2015, next year after the flood. For 57.1% of trees which released growth in 2015, this was the most significant increase of ring width during their whole lives (in whole tree-ring chronologies), and for 23.8% of trees which reduced growth in 2015, this was the most significant decrease of ring width. Tree-ring reductions in next year after the 2014 flood resulted from environmental stress related to burying stems with alluvia deposited during the flood. Fresh sediments cut off air access from the root system, and for some of the sampled trees, this stress was strong enough to control their radial growth. Growth releases that follow the 2014 flood are a record of trees benefitting from a sudden supply of water, a rare opportunity in dry study area, where water is usually scarce. The study demonstrates that floods in high mountains of arid zone can cause dual, opposite growth reaction of affected trees. Such dual record is characteristic for environmental impulses, which exert stress on trees, but, at the same time, improve other conditions of tree growth. Environmental events that cause simultaneous positive and negative reactions among a population of trees, like studied flood, can easily be overlooked in chronologies based on average widths of tree rings each year. For trees affected during studied flood arithmetical mean of ring widths in 2015 is average and does not stand out from arithmetical means for other years. However, when analysed in detail, the year 2015 is significantly different from other years, as is demonstrated by high values of dispersion indicators (standard deviation and coefficient of variation) calculated for all sampled trees. This study demonstrates that following the standard procedure (developing tree-ring chronology from average ring widths) is not a reliable solution for reconstructions of environmental impulses which cause dual, opposite reaction among sampled trees. Even strong events of this type will not be emphasised in standard chronologies, which can lead to underestimating frequency and magnitude of processes and, in the case of floods, to underestimating hazard and risk.