2nd WP8 Meeting on Cross-Scale Interaction

16-17 February 2017, Potsdam (D)

Ariane Walz (Univ. of Potsdam), Elisa Palazzi (CNR), Emiliana Valentini (ISPRA), Arnon Karnieli (BGU) and Ed Wheatcroft (LSE)

In February 16-17 2017, many of the WP8 core group made it again to Potsdam for our second Work Package Meeting! Here, we used the opportunity to highlight some of the latest developments:

Reconstructing past changes. We reconstruct past changes for all ECOPOTENTIAL PAs to increase the comparability across ECOPOTENTIAL sites and to bridge from individual single storylines to knowledge that might be of interest for PA management all over Europe. We address therefore variables that are commonly interesting to all mountain, arid or coastal/marine ecosystems. They will be derived from Remote Sensing data (mainly LANDSAT), from gridded climate data (E-OBS, or re-analyses data such as ERAInterim-Land) and if possible from in-situ data (still discussed with WP5). To keep the extra work to a minimum, the study takes up many remote sensing data already requested by partners. A number of partners across different work packages have started to engage in the endeavor, and if you are interested, please feel free to get in touch! Future simulations. In close collaboration with WP6/Ecosystem Modelling, we started a concerted action on future simulations for ten PAs. Despite differences in the ecosystem models used and in the ecosystems investigated, we all follow the same modelling protocol (as defined at the Pisa Workshop on Future Projection in September 2016) and use the climate data downscaled by Elisa Palazzi and Silvia Terzago (CNR-ISAC, Torino). First of all, the downscaling strategy has been illustrated and summarized at the Workshop: It includes the use of specific temperature and precipitation downscaling methods, both taking the orography into account, to be applied to the outputs of the Regional Climate Model RCA4 from the CORDEX model ensemble, driven by five different Global Climate Models from the CMIP5 experiment. RCA4 is the only model run at 0.11 degrees (~12 km) spatial resolution providing data with a sub-daily temporal resolution (3 h). Since for both precipitation and temperature we perform a spatial downscaling only (up to around 1 km, or slightly less), the use of that specific high-temporal resolution RCM would allow to achieve sub-daily temporal resolution in the final output, too, which was one of the requests of ECOPOTENTIAL partners who asked for downscaled data. The strategy also includes bias-correction of the RCM data using the EOBS datasets as the reference before downscaling is applied. After resuming the strategy, the downscaled data already delivered, under delivery and in production were presented and those still to be produced were listed. New requests from the participants to the workshop are still possible! An example of precipitation downscaling over the Gran Paradiso PA is shown in Fig. 1. The original field (climatological annual average for the period 1970-2005) from RCA4 driven by the EC-Earth GCM is at 11 km spatial resolution; the downscaled field at 1 km spatial resolution. Fig.2 shows another example in which temperature time series from the bias-corrected RCM (yellow) and the bias-corrected + downscaled RCM (black) are extracted in correspondence to the location of two measurement stations whose time series are also reported in the figure (Serru in the left, Teleccio in the right; red lines). For the model, the historical and future (RCP4.5 and RCP8.5 scenarios) temperature time series are displayed.

Figure 1. Multiannual mean, averaged over the period 1979-2005, of the precipitation data from the Regional Climate Model RCA4 at 11 km resolution (left) and downscaled at 1 km resolution (right).©Elisa Palazzi.
Figure 2. Temperature time-series for two locations in the Gran Paradiso National Park PA, Serru (left) and Teleccio (right). Measured data are in red, bias-corrected RCM data are in yellow, bias-corrected plus downscaled RCM data for historical and future conditions are in black /gray. ©Elisa Palazzi.

Multi-scale analysis. Arnon Karnieli presented first results from a multi-scale study on the HarHa Negev. Combing LANDSAT and RapidEye data allows to cover both large-extend and high-resolution in space and time in this study.

Figure 3. Reconstructed changes in land cover in the HarHa Negev based on LANDSAT (©Arnon Karnieli).
Figure 4. Four patterns of Sea Surface Temperature based on Empirical Orthogonal Function maps (Source: ISPRA – Institute for Environmental Protection and Research. Elaborated by Emiliana Valentini and Federico Filliponi).

Cross-scale analysis for the Mediterranean. Emiliana Valentini and Federico Filipponi presented results for a pan-Mediterranean reconstruction of daily Sea Surface Temperature (SST) from 1982 to 2015 and monthly Chlorophyll-a (Chl-a) concentration from 2008 to 2014, demonstrating the value of products provided by Copernicus services. They are using these data to investigate variability across spatial and temporal scales to find ‘tipping points’ in spatial patterns and temporal trends. These first results triggered inspiring discussions during the workshop!

WP8 links up to Essential Variables. Carlos Guerra informed us about the latest results of the bottom-up identification of Essential Variables as an overlap of variables addressed in the storylines. To our great relieve, it turned out that about two thirds of the variables will most likely be also addressed in a number of WP8 activities.

NOT TO MISS: Training workshop in August / September 2017. Lenny Smith and Ed Wheatcroft from London School of Economics will invite you to a hands-on workshop to estimate and increase the forecasting power of your model. More information to come!