MAIDENiso: a mechanistic approach to the reconstruction of past climate from tree chronologies

The ecophysiological forest model MAIDENiso (Modeling and Analysis In + isotopes) uses a set of mechanistic rules to simulate the production, allocation and growth of virtual trees. MAIDENiso is adapted to the boreal tree species Picea mariana Mill. (Black spruce), but lacks a hydrological module adapted for boreal meteorological conditions. With the recent addition of a snow/ice module, MAIDENiso is now capable of realistically simulating snow cover and discharge in high latitude regions, while at the same time capturing climate-sensitive processes such as the enrichment of heavy water isotopes due to snow sublimation. The more realistic outputs of the model can be compared to tree ring records (ring widths and stable isotopes). This allows us to use an inversion algorithm (based on a Metropolis Hastings random walk) to estimate past hydroclimate conditions that are in line with physiological and hydrological processes of high boreal regions. We apply this methodology to a millennial chronology of tree ring width and cellulose isotopes from sub-fossil tree remains in North-Quebec, and produce an updated hydroclimate reconstruction of the last 1000 years in this region.