Palaeohydrology, fires and vegetation succession in the southern Baltic during the last 7500 years reconstructed from a raised bog based on multi-proxy data

We present the first 7500 yr long multi-proxy record from a raised bog located at the southern Baltic coast, Poland. Testate amoebae, plant macrofossils, pollen and microscopic charcoal were used to reconstruct environmental changes in Pomerania (northern Poland, Kaszuby Lakeland) from a 7-m thick peat archive of Stążki bog dated 5500 BC–AD 1250. We obtained a record of proxies representing different spatial scales: regional vegetation changed simultaneously with local vegetation, and testate amoebae showed a pattern of change similar to that of pollen and plant macrofossils. On the basis of the combined proxies, we distinguished three hydroclimatic stages: moist conditions 5500–3450 BC, drier conditions with regionally increased fires up to 600 BC, and again moist conditions from 600 BC onward. During the drier interval, a first climatic shift to wetter conditions at 1700 BC is indicated by regional pollen as the replacement of Corylus by Carpinus, and locally by, e.g., the increase of Hyalosphenia elegans and mire plants such as Sphagnum sec. Cuspidata. Furthermore, we observed a correlation since 600 BC among the re-expansion of Carpinus (after a sudden decline ca. 950 BC), increased peat accumulation, increase of Hyalosphenia species, and fewer fires, suggesting lower evapotranspiration and a stable high water table in the bog. Fagus started to expand after AD 810 gradually replacing Carpinus, which was possibly due to a gradually more oceanic climate, though we cannot exclude human impact on the forests. Peat accumulation, determined by radiocarbon dating, varied with bog surface wetness. The hydroclimatic phases found in Stążki peatland are similar to moisture changes recorded in other sites from Poland and Europe. This is the first detailed record of hydroclimatic change during the Holocene in the southern Baltic region, so it forms a reference site for further studies on other southern Baltic bogs that are in progress.

Multi-Model Data Fusion to Improve an Early Warning System for Hypo-/Hyperglycemic Events

Correct predictions of future blood glucose levels in individuals with Type 1 Diabetes (T1D) can be used to provide early warning of upcoming hypo-/hyperglycemic events and thus to improve the patient's safety. To increase prediction accuracy and efficiency, various approaches have been proposed which combine multiple predictors to produce superior results compared to single predictors. Three methods for model fusion are presented and comparatively assessed. Data from 23 T1D subjects under sensor-augmented pump (SAP) therapy were used in two adaptive data-driven models (an autoregressive model with output correction - cARX, and a recurrent neural network - RNN). Data fusion techniques based on i) Dempster-Shafer Evidential Theory (DST), ii) Genetic Algorithms (GA), and iii) Genetic Programming (GP) were used to merge the complimentary performances of the prediction models. The fused output is used in a warning algorithm to issue alarms of upcoming hypo-/hyperglycemic events. The fusion schemes showed improved performance with lower root mean square errors, lower time lags, and higher correlation. In the warning algorithm, median daily false alarms (DFA) of 0.25%, and 100% correct alarms (CA) were obtained for both event types. The detection times (DT) before occurrence of events were 13.0 and 12.1 min respectively for hypo-/hyperglycemic events. Compared to the cARX and RNN models, and a linear fusion of the two, the proposed fusion schemes represents a significant improvement.

Index Tracking Using Data-Mining Techniques and Mixed-Binary Linear Programming

Index tracking has become one of the most common strategies in asset management. The index-tracking problem consists of constructing a portfolio that replicates the future performance of an index by including only a subset of the index constituents in the portfolio. Finding the most representative subset is challenging when the number of stocks in the index is large. We introduce a new three-stage approach that at first identifies promising subsets by employing data-mining techniques, then determines the stock weights in the subsets using mixed-binary linear programming, and finally evaluates the subsets based on cross validation. The best subset is returned as the tracking portfolio. Our approach outperforms state-of-the-art methods in terms of out-of-sample performance and running times.

A multi-proxy approach for revealing recent climatic changes in the Russian Altai

For the first time we present a multi-proxy data set for the Russian Altai, consisting of Siberian larch tree-ring width (TRW), latewood density (MXD), δ13C and δ18O in cellulose chronologies obtained for the period 1779–2007 and cell wall thickness (CWT) for 1900–2008. All of these parameters agree well between each other in the high-frequency variability, while the low-frequency climate information shows systematic differences. The correlation analysis with temperature and precipitation data from the closest weather station and gridded data revealed that annual TRW, MXD, CWT, and δ13C data contain a strong summer temperature signal, while δ18O in cellulose represents a mixed summer and winter temperature and precipitation signal. The temperature and precipitation reconstructions from the Belukha ice core and Teletskoe lake sediments were used to investigate the correspondence of different independent proxies. Low frequency patterns in TRW and δ13C chronologies are consistent with temperature reconstructions from nearby Belukha ice core and Teletskoe lake sediments showing a pronounced warming trend in the last century. Their combination could be used for the regional temperature reconstruction. The long-term δ18O trend agrees with the precipitation reconstruction from the Teletskoe lake sediment indicating more humid conditions during the twentieth century. Therefore, these two proxies could be combined for the precipitation reconstruction.

Climate and human impact during the past 2000 years as recorded in the Lagunas de Yala, Jujuy, northwestern Argentina

210Pb, 137Cs and 14C dated sediments of two late Holocene landslide lakes in the Provincial Park Lagunas de Yala (Laguna Rodeo, Laguna Comedero, 24°06′S, 65°30′W, 2100 m asl, northwestern Argentina) reveal a high-resolution multi-proxy data set of climate change and human impact for the past ca. 2000 years. Comparison of the lake sediment data set for the 20th century (sediment mass accumulation rates MARs, pollen spectra, nutrient and charcoal fluxes) with independent dendroecological data from the catchment (fire scars, tree growth) and long regional precipitation series (from 1934 onwards) show that (1) the lake sediment data set is internally highly consistent and compares well with independent data sets, (2) the chronology of the sediment is reliable, (3) large fires (1940s, 1983/1984–1989) as documented in the local fire scar frequency are recorded in the charcoal flux to the lake sediments and coincide with low wet-season precipitation rates (e.g., 1940s, 1983/1984) and/or high interannual precipitation variability (late 1940s), and (4) the regional increase in precipitation after 1970 is recorded in an increase in the MARs (L. Rodeo from 100 to 390 mg cm−2 yr−1) and in an increase in fern spores reflecting wet vegetation. The most significant change in MARs and nutrient fluxes (Corg and P) of the past 2000 years is observed with the transition from the Inca Empire to the Spanish Conquest around 1600 AD. Compared with the pre-17th century conditions, MARs increased by a factor of ca. 5 to >8 (to 800 +130, −280 mg cm−2 yr−1), PO4 fluxes increased by a factor of 7, and Corg fluxes by a factor of 10.5 for the time between 1640 and 1930 AD. 17th to 19th century MARs and nutrient fluxes also exceed 20th century values. Excess Pb deposition as indicated by a significant increase in Pb/Zr and Pb/Rb ratios in the sediments after the 1950s coincides with a rapid expansion of the regional mining industry. Excess Pb is interpreted as atmospheric deposition and direct human impact due to Pb smelting.

Evaluation of Three State-of-the-Art Classifiers for Recognition of Activities of Daily Living from Smart Home Ambient Data

Smart homes for the aging population have recently started attracting the attention of the research community. The "health state" of smart homes is comprised of many different levels; starting with the physical health of citizens, it also includes longer-term health norms and outcomes, as well as the arena of positive behavior changes. One of the problems of interest is to monitor the activities of daily living (ADL) of the elderly, aiming at their protection and well-being. For this purpose, we installed passive infrared (PIR) sensors to detect motion in a specific area inside a smart apartment and used them to collect a set of ADL. In a novel approach, we describe a technology that allows the ground truth collected in one smart home to train activity recognition systems for other smart homes. We asked the users to label all instances of all ADL only once and subsequently applied data mining techniques to cluster in-home sensor firings. Each cluster would therefore represent the instances of the same activity. Once the clusters were associated to their corresponding activities, our system was able to recognize future activities. To improve the activity recognition accuracy, our system preprocessed raw sensor data by identifying overlapping activities. To evaluate the recognition performance from a 200-day dataset, we implemented three different active learning classification algorithms and compared their performance: naive Bayesian (NB), support vector machine (SVM) and random forest (RF). Based on our results, the RF classifier recognized activities with an average specificity of 96.53%, a sensitivity of 68.49%, a precision of 74.41% and an F-measure of 71.33%, outperforming both the NB and SVM classifiers. Further clustering markedly improved the results of the RF classifier. An activity recognition system based on PIR sensors in conjunction with a clustering classification approach was able to detect ADL from datasets collected from different homes. Thus, our PIR-based smart home technology could improve care and provide valuable information to better understand the functioning of our societies, as well as to inform both individual and collective action in a smart city scenario.