Description: The second Data Science Symposium at AWI gathered several data science related talks from AWI, GEOMAR and HZG.

Description: The focus of this research has been on detecting changes in lake areas, vegetation, land surface temperatures, and the area covered by snow, using data from remote sensing. The study area covers the main (central) part of the Lena catchment in the Yakutia region of Siberia (Russia), extending from east of Yakutsk to the central Siberian Plateau, and from the southern Lena River to north of the Vilyui River. Approximately 90% of the area is underlain by continuous permafrost. Remote sensing products were used to analyze changes in water bodies,land surface temperature (LST), and leaf area index (LAI), as well as the occurrence and extent of forest fires,and the area and duration of snow cover. The remote sensing analyses (for LST, snow cover, LAI, and fire) were based on MODIS–derived NASA products for 2000 to 2011. Changes in water bodies were calculated from two mosaics of (USGS) Landsat high resolution (30 m) satellite images from 2002 and 2009. Within the study area’s 315,000 km2 the total area covered by lakes increased by 17.5% between 2002 and 2009, but this increase varied in different parts of the study area, ranging between 11% and 42%. The land surface temperatures showed a consistent warming trend, with an average increase of about 0.12�C/year. The average rate of warming during the April-May transition period was 0.15�C/year and 0.19 �C/year in the September-October period, but ranged up to 0.45�C/year in some areas during April-May. Regional differences in the rates of land surface temperature change, and possible reasons for the temperature changes, are discussed with respect to changes in the land cover. Our analysis of a broad spectrum of variables over the study area suggests that the spring warming trend is very likely to be due to changes in the area covered by snow. The warming trend observed in fall does not, however,appear to be directly related to any changes in the area of snow cover, or to the atmospheric conditions, or to the proportion of the land surface that is covered by water (i.e. to wetting and drying). These results suggest a complex interplay between different mechanisms affecting the land cover and land surface temperatures that warrants further investigation, possibly making use of higher resolution satellite data together with local and regional modeling, and taking into account the influence of lakes on the regional energy exchange. Supplementary data (original data, digitized version of the maps, metadata) are archived under PANGAEA (http://dx.doi.org/10.1594/PANGAEA.855124).

Description: The focus of this research has been on detecting changes in lake areas, vegetation, land surface temperatures, and the area covered by snow, using data from remote sensing. The study area covers the main (central) part of the Lena River catchment in the Yakutia region of Siberia (Russia), extending from east of Yakutsk to the central Siberian Plateau, and from the southern Lena River to north of the Vilyui River. Approximately 90% of the area is underlain by continuous permafrost. Remote sensing products were used to analyze changes in water bodies, land surface temperature (LST), and leaf area index (LAI), as well as the occurrence and extent of forest fires, and the area and duration of snow cover. The remote sensing analyses (for LST, snow cover, LAI, and fire) were based on MODIS–derived NASA products (250–1000 m) for 2000 to 2011. Changes in water bodies were calculated from two mosaics of (USGS) Landsat (30 m) satellite images from 2002 and 2009. Within the study area's 315,000 km2 the total area covered by lakes increased by 17.9% between 2002 and 2009, but this increase varied in different parts of the study area, ranging between 11% and 42%. The land surface temperatures showed a consistent warming trend, with an average increase of about 0.12 °C/year. The average rate of warming during the April–May transition period was 0.17 °C/year and 0.19 °C/year in the September–October period, but ranged up to 0.49 °C/year during September–October. Regional differences in the rates of land surface temperature change, and possible reasons for the temperature changes, are discussed with respect to changes in the land cover. Our analysis of a broad spectrum of variables over the study area suggests that the spring warming trend is very likely to be due to changes in the area covered by snow. The warming trend observed in fall does not, however, appear to be directly related to any changes in the area of snow cover, or to the atmospheric conditions, or to the proportion of the land surface that is covered by water (i.e., to wetting and drying). Supplementary data (original data, digitized version of the maps, metadata) are archived under PANGAEA (http://dx.doi.org/10.1594/PANGAEA.855124).

Description: Arctic lowlands are characterized by large numbers of small waterbodies, which are known to affect surface energy budgets and the global carbon cycle. Statistical analysis of their size distributions has been hindered by the shortage of observations at sufficiently high spatial resolutions. This situation has now changed with the high-resolution (〈5 m) circum-Arctic Permafrost Region Pond and Lake (PeRL) database recently becoming available. We have used this database to make the first consistent, high-resolution estimation of Arctic waterbody size distributions, with surface areas ranging from 0.0001 km2 (100 m2) to 1 km2. We found that the size distributions varied greatly across the thirty study regions investigated and that there was no single universal size distribution function (including power-law distribution functions) appropriate across all of the study regions. We did, however, find close relationships between the statistical moments (mean, variance, and skewness) of the waterbody size distributions from different study regions. Specifically, we found that the spatial variance increased linearly with mean waterbody size (R2 = 0.97, p 〈 2.2e-16) and that the skewness decreased approximately hyperbolically. We have demonstrated that these relationships (1) hold across the 30 Arctic study regions covering a variety of (bio)climatic and permafrost zones, (2) hold over time in two of these study regions for which multi-decadal satellite imagery is available, and (3) can be reproduced by simulating rising water levels in a high-resolution digital elevation model. The consistent spatial and temporal relationships between the statistical moments of the waterbody size distributions underscore the dominance of topographic controls in lowland permafrost areas. These results provide motivation for further analyses of the factors involved in waterbody development and spatial distribution and for investigations into the possibility of using statistical moments to predict future hydrologic dynamics in the Arctic.

Description: Permafrost acts as an impermeable subsurface in Arctic lowland landscapes. This hydrological barrier results in carbon-rich, water-saturated soils as well as many ponds and lakes. The rapidly warming Arctic climate very likely will affect the surface inundation in Arctic lowlands due to changes in precipitation, evapotranspiration, and permafrost degradation. Drying and wetting of the surface may occur in different regions and potentially alter the exchange of energy and carbon between the surface and the atmosphere. With increased permafrost thaw, for example, water may drain to deeper soil layers or drainage maybe enhanced due to newly forming drainage networks. Melting ground ice and subsequent inundation, on the other hand, may enhance formation of new ponds and wet areas. The current distribution of ponds and lakes in the Arctic is the result of complex interactions between climate, ground ice volume, topography, age and sediment characteristics. Because lake formation and growth processes occur at spatial scales orders of magnitude below those of the resolution for global or pan-arctic models land surface models, statistical representations of lake size distributions and other properties to inform such processes in future models are needed that can be related to macroscopic landcape properties. This study proposes basic observationally-constrained relationships to enhance the modeling of future Arctic surface inundation. We mapped ponds and lakes in 21 circum-arctic sites representing different permafrost-soil landscapes, i.e., physiographic regions with similar surface geology, regional climate, and biomes. We used high-resolution optical and radar satellite imagery with spatial resolutions of 4 m or better to create detailed water body maps and derive representative probability density functions (PDF). PDFs of ponds and lakes vary little within the same ecoregion. Significant differences, however, do occur between landscapes. We used regional permafrost-soil landscape maps of Alaska, Canada, and Siberia to upscale the water body distributions to the circum-arctic. We here present regional distribution parameters, i.e. pond and lake fractions as well as PDF moments (mean surface area, standard deviation, and skewness) and their uncertainties. Younger landscapes, that developed in the early Holocene exhibit very skewed water body distributions. These landscapes are dominated by many ponds and feature only very few large lakes. Older landscapes, on the other hand, show more larger lakes but also a higher variability in pond and lake size. For lakes smaller than 5*10⁵ m², PDFs change in a regular fashion across all sites: Relationships between mean surface area and standard deviation show a linear behaviour whereas the correlation between mean and skewness log-normal. We hypothesize that these relationships are an expression of pond and lake growth and/or lake formation in the landscapes and discuss the potential of the observed patterns to improve predictions of future distributions of Arctic ponds and lakes.

Description: Most of the world's permafrost is located in the Arctic, where its frozen organic carbon content makes it a potentially important influence on the global climate system. The Arctic climate appears to be changing more rapidly than the lower latitudes, but observational data density in the region is low. Permafrost thaw and carbon release into the atmosphere, as well as snow cover changes, are positive feedback mechanisms that have the potential for climate warming. It is therefore particularly important to understand the links between the energy balance, which can vary rapidly over hourly to annual timescales, and permafrost conditions, which changes slowly on decadal to centennial timescales. This requires long-term observational data such as that available from the Samoylov research site in northern Siberia, where meteorological parameters, energy balance, and subsurface observations have been recorded since 1998. This paper presents the temporal data set produced between 2002 and 2017, explaining the instrumentation, calibration, processing, and data quality control. Furthermore, we present a merged data set of the parameters, which were measured from 1998 onwards. Additional data include a high-resolution digital terrain model (DTM) obtained from terrestrial lidar laser scanning. Since the data provide observations of temporally variable parameters that influence energy fluxes between permafrost, active-layer soils, and the atmosphere (such as snow depth and soil moisture content), they are suitable for calibrating and quantifying the dynamics of permafrost as a component in earth system models. The data also include soil properties beneath different microtopographic features (a polygon centre, a rim, a slope, and a trough), yielding much-needed information on landscape heterogeneity for use in land surface modelling. For the record from 1998 to 2017, the average mean annual air temperature was −12.3 ∘C, with mean monthly temperature of the warmest month (July) recorded as 9.5 ∘C and for the coldest month (February) −32.7 ∘C. The average annual rainfall was 169 mm. The depth of zero annual amplitude is at 20.75 m. At this depth, the temperature has increased from −9.1 ∘C in 2006 to −7.7 ∘C in 2017. The presented data are freely available through the PANGAEA (https://doi.org/10.1594/PANGAEA.891142) and Zenodo (https://zenodo.org/record/2223709, last access: 6 February 2019) websites.

Description: Arctic ponds, i. e. water bodies with a surface area equal to or smaller than 10⁴ m² (1 ha), are currently not inventoried on a circum-arctic scale. However, they are a key element of the water, energy, and carbon balance and abundant in Arctic permafrost lowlands. Ponds and lakes have been subject to both wetting and drying in a warming climate yet studies remain ambivalent regarding the causes of these changes. Goals of this study are to (i) investigate the variability of water body size distributions as a function of landscape characteristics, and (ii) assess the vulnerability of water bodies in different landscapes to scenarios of wetting and drying. Ponds and lakes were mapped from high-resolution aerial and satellite imagery with resolutions of 4 m or better in 14 regions in Alaska, Canada, and Siberia covering a total area of ca. 1.6*104 km². Whereas lake distributions are similar, pond distributions in our study regions vary significantly with the area-normalized number of ponds differing up to 3 orders of magnitude. Landscape characteristics that may explain the current water body distributions include climate (eg., precipitation, evapotranspiration, temperature), permafrost (eg., ground ice content, maximum thaw depth) and terrain characteristics (eg., topography, glaciation, landscape age) which we derive from in situ, remote sensing and modeling data sources. Multivariate regression analysis are used to relate landscape characteristics to distribution parameters. This study for the first time allows to quantify the circum-arctic variability of pond distribution. The current maps are the start of a high-resolution circum-arctic water body inventory and present a baseline for future surface inundation mapping and modelling. We present representative regional probability density functions (pdf) and assess the potential to upscale pdfs using spatial landscape characteristics. We then discuss the vulnerability of water bodies to wetting or drying based on the distribution parameters, their correlation with landscape characteristics and the likeliness of both to change in different future climate scenarios.