Data for: Late Holocene ecological shifts and chironomid-inferred summer temperature changes reconstructed from lake Uddelermeer, the Netherlands

In the paper accompanying this dataset a Late-Holocene chironomid-inferred, July-air temperature (CI-T) record is presented. The was data collected from a core taken from lake Uddelermeer, a pingo remnant located in the central Netherlands (52º14’48’’N; 5º45’40’’E; 24-27 m a.s.l.). The core interval dating from 2500 to 400 cal. yr. BP was analysed on multi-decadal resolution for pollen, spores and NPPs, and chironomid head capsules. The dataset contains three files: • The dataset Pollen count_figure 3 contains pollen count data for the core segment 100-500 cm from lake Uddelermeer shown in figure 3 in the accompanying article. Pollen and spores were identified using (Beug 2004), (Moore et al. 1991) and (Punt et al. 1976; 1980; 1981; 1984; 1988; 1991; 1995; 2003; 2009). NPPs were identified using (1972; 1978; van Geel et al. 2003). • The dataset Chironomid count_figure 4 contains chironomid head capsule count for the core segment 126 - 549 cm from lake Uddelermeer shown in figure 4 of the accompanying article. This file also contains data from samples previously analysed by Stefan Engels and published in (Engels et al. 2016). Chironomid head capsules were identified following keys by (Wiederholm 1983; Moller Pillot 1984a; Moller Pillot 1984b; Rieradevall and Brooks 2001; Brooks et al. 2007). • The dataset temperature reconstruction_figure 5 contains a chironomid inferred temperature reconstruction with relevant statistics shown in figure 5 of the accompanying article. The chironomid head capsule data was transferred to a July-air temperature reconstruction using the Norwegian-Swiss chironomid-temperature inference model based on a modern calibration dataset from 274 lakes from Norway and the Alpine regions and spans a July air temperature range from 3.5 to 18.4°C (Heiri et al. 2011). For the model, a two component weighted averaging partial least-squares (WA-PLS) regression was used with a bootstrapped root mean square error of prediction of 1.40°C after outlier deletion (Heiri et al. 2011). The closest modern analogue was assessed based on squared Chi2 distances using the program C2 (Juggins 2003). Fossil samples with a distance to the closest analogue larger than the 2nd and 5th percentile of the distances of the modern samples in the calibration dataset were classified as having ‘no close’ and ‘no good’ analogue, respectively. Goodness-of-fit to temperature was assessed using a Canonical Correspondence Analysis (CCA) of the modern samples with temperature as the only constraining variable using CANOCO for Windows version 4.51 (ter Braak and Šmilauer 2012). In this analysis, the fossil samples were added passively. Fossil samples with a residual distance exceeding the 90th and 95th percentiles of the residual distances of the modern samples were identified as having a ‘poor’ and ‘very’ poor fit with temperature, respectively. Numerical cut-off thresholds for the closest modern analogue and goodness-of-fit are provided in the spreadsheet