Biodiversity / Aquatic Ecosystems

High latitude wetlands are important for understanding climate change risks because these environments sink carbon and emit methane. West Siberia Plain is the biggest peatland area in Eurasia and is situated in the high latitudes experiencing enhanced rate of climate change. Fine scale heterogeneity of wetland landscapes pose challenges for producing the greenhouse gas flux inventories based on point observations. A number of peatland maps of the West Siberia was developed in 1970s, but their accuracy is limited. To reduce uncertainties at the regional scale, we mapped wetlands in the taiga zone of West Siberia (WS) on a scene-by-scene basis using a supervised classification of Landsat imagery. The training datasets were based on high-resolution images and ground truth data collected for more than 10 years of fieldwork in West Siberia summarized in an extensive dataset of botanical descriptions, field photos, pH and electrical conductivity data from 40 study areas. Classification scheme was oriented on methane inventory applications and included seven wetland ecosystem types composing nine wetland complexes in different proportions. Accuracy assessment based on 1082 validation polygons of 10×10 pixels indicated an overall map accuracy of 79%. The total area of the wetlands and water bodies was estimated to be 70.8 Mha or 26% of the domain; three-quarters of wetlands locate in taiga zone (55°N-65°N). Methane emission rates vary strongly among the wetland ecosystems. While the highest amounts are emitted by waterlogged hollows followed by fens, elevated environments as forested bogs and ridges emit at the lowest rates. According to our estimates, the later account for only 2% of the regional total emissions while occupying almost 40% of the wetland area. The oligotrophic and waterlogged hollows (the latter includes small lakes with area < 60×60 m2) as parts of patterned bogs cover a third of the mire territory accounting for 60% of the total CH4 emission. Large (> 60×60 m2) lakes and rivers  are widespread over taiga mire systems occupying 12% of the territory and contributing only 3% of CH4 flux to the atmosphere. The rest of the territory is mainly occupied by different fens, poor fens and swamps (14%) contributing 35% of the total CH4 flux. Concerning zonal distribution, taiga contributes 85% to regional methane flux and tundra only 8%, however ebullition in lakes was not directly measured. Applying the new map resulted in total methane emissions of 4.62 TgCH4/yr, which is 72% higher than the estimate based on the same emission dataset and a map by Peregon et al. (2009). The revision resulted from the changes in fractional coverages of methane emitting ecosystems. New Landsat-based map of WS’s taiga wetlands provides a useful benchmark for validation of coarse-resolution global land cover products and wetland datasets in high latitudes.

Monitoring of long-term land-use and land-cover change patterns may be biased by seasonal changes of different surface properties (e.g. hydrology, phenology, etc.) which become even more prominent in highly dynamic ecosystems such as wetlands (Crews-Meyer, 2008; McClearly, Crews-Meyer and Young 2008; Dronova et al. 2011). These surface dynamics produce transitional states and fine-scale mixtures of classes that may hinder classifications and long-term change detection. Dronova et al. (2015) proposed the term “Dynamic Cover Types" (DCT) to refer to such areas of regimes of periodic or seasonal change. Examples of DCT in the context of wetlands would be seasonally inundated forests, temporal water bodies and waterways, or harvests of reeds and crops such as rice.

We assess the spatio-temporal extent of DCT in two study sites; The Camargue, a large coastal wetland in Southern France, and the Lagoon of Fuente de Piedra, a small wetland in Southern Spain. For that we use a multitemporal change detection procedure for polarimetric SAR imagery based on the Complex Wishart distribution developed recently by Conradsen et al (2015), (to be published) and an innovative open source software implementation which makes use of Ipython Notebooks and Docker containers (http://mortcanty.github.io/SARDocker/). The procedure carries out a series of change detection processing routines for the whole time series with a desired significance level. It uses multilook, geocoded and terrain corrected intensity images in C2 matrix. These were generated in the Sentinel Application Platform (SNAP) using 12 Sentinel-1 images (Interferometric Wide, Single Look Complex and cross-polarized) with a monthly resolution.

The methodology proposed here for change detection is relatively easy to use and utilizes only open source and free data. It enables an operational monitoring service of short-term change detection. No calibration or validation needed, only interpretation of changes using local knowledge. This has important implications for operational standardized monitoring service such as the ones developed in the –Satellite-based Wetland Observation Service (SWOS) Horizon 2020 project.

Besides its easiness to use, this methodology has other important advantages: First, the fine spatial and temporal resolutions of Sentinel-1 SAR data allow us to detect short-time changes for a complete water year regardless of the cloud cover. Second, change detection methods based on classification are affected by classification errors, whose probability of occurrence increases in dynamic and transitional landscapes (Powell et al. 2003). Our approach does not rely on classification and thus is free from such errors. Third, DCT are complex landscapes that often give rise to unique species assemblages (Parrot & Meyer 2012; Watson et. al 2014), and knowing their spatio-temporal extent will assist in biodiversity management. Fourth, annual stable features can be identified and used for training areas, which may facilitate the classification process and improve accuracies. And fifth, estimating the spatio-temporal extent of DCT might shed some light on the wide array of options in classification methodologies available and their different results (Object vs. Pixel based, Support Vector Machines, Random Forest Classifiers, and other algorithms).

Wetlands and land surface waters play a crucial role in the global water and biogeochimal cycles, on the management of water resources and in the climate. In the climate change context, it essential to better describe and understand each component of the water cycle. Nowadays, our understanding of the open water over land is still incomplete as it plays a major role in the water cycle. Since the 80's, remote sensing techniques provided quantitative estimates of open water over land and appear to be a perfect tool to monitor natural and anthropogenic evolution of this variable over the globe. Most of the studies on this topic use visible, infrared and active microwave to observe open water over land. In this topic, authors use the Soil Moisture and Ocean Salinity (SMOS) passive microwave sensor in L-band, launch in 2009 by the European Space Agency (ESA). At this frequency, the signal is highly sensitive to standing water above the ground. At L-band, the signal is expected to penetrate deeper in vegetation than signal in other frequency, such as visible and infrared. This asset permits to L-band signal to be more sensitive to open water under dense vegetation. Moreover SMOS has a high temporal resolution (3-days).

In this study, concerns the Amazon and Congo basins where an important part of the basin is covered by dense vegetation. A first order radiative model had been built to obtain surface open water over land from the brightness temperature (physical signal) measured by SMOS. At coarse resolution (25 km x 25 km), a 2-days maps of water fraction is produced over the Amazon and Congo basin. These new SMOS products are validated against altimeter data (Jason-2) over the Rio Negro river. It was found that the water fraction estimated by SMOS was highly correlated with water levels measured by Jason-2 (R > 0.98). They are also compared to the static land cover map such as Ecoclimap and the International Geosphere-Biosphere Program (IGBP). Also the maps exhibit a phase shift of three months between the South and the North of the Amazon basin in consistency with the climatic regime across the basin.

A dense time series of the Alos PALSAR WB radar data was used for the mapping of the wetlands at the continental Colombian territory. Radar data processing, ortho reactification and calibration was done in SarVision using in house created software and Gamma software, resulting in high quality 28 intercalibrated mosaics, with an intercalibration error not higher than 0.3 dB.  Each mosaic was classified using a range of values, evaluated for each mosaic, in order to differentiate two types of flooding.1) open sky flooding and 2) flooding under the canopy. Ecosystem knowledge was integrated to create post-processing rules to correct the flooding maps from errors created by confusion in the scattering mechanims.. An structural vegetation map (created with Alos PALSAR FB data) was used to correct the classifications. The information of the 28 corected classified mosaics is integrated into a flooding frequency map, used as the main input for the mapping of the wetlands in Colombia. Different types of wetlands can be delineated when integrated with the vegetation structure information.

The use of the ecosystem approach increases the overall classification accuracy in more than 20 %. Validation was done using field data and high resolution images, over more than 8000 locations, along the different ecosystems,  Most errors are found in the open flooding ecosystems occuring in the savanas and in areas of  low vegetation. errors can be explained with the behavior of the radar waves and description of main scattering mechanisms occuring in this ecosystems. the final map can be regardes as a baseline for the study of the floodings in Colombia, since it incorporates information of five years ( 2007-2011)including extreme El Niño- La Niña events.

Over almost ten years, the European Space Agency operated the largest Earth observation satellite built to date, ENVISAT. Its instruments provided optical, thermal and altimeter observations, and allowed for retrieval accuracies not previously achieved by spaceborne remote sensing. Using these observations, the European Space Agency launched the DUE project DIVERSITY II to create a comprehensive, publicly available database consisting of water quality and quantity parameters for more than 300 lakes. The database is now publicly available at www.diversity2.info, together with corresponding documentation, several case studies and a collection of Python functions for display tasks and spatio-temporal analyses. The Diversity II inland water database is designed to meet the requirements of the aquatic biodiversity community, but represents beyond this purpose the first globally consistent and reproducible knowledge basis of its kind.

The database consists of water composition, level and temperature variations. Water composition parameters are retrieved from MERIS full resolution imagery and compiled as level 3 GeoTIFF products of monthly, yearly and 9-year aggregation, using a dedicated processing chain for valid pixel identification, atmospheric correction, water type classification and constituent retrieval. The parameters included in each product are chlorophyll-a, suspended matter and CDOM concentration, turbidity, cyanobacteria and floating matter probability. Water level time series are adapted as individual observations from the Hydroweb database (Cretaux et al., 2011), and available as tables and time series figures for about one third of all lakes selected in Diversity II. Monthly lake surface water temperature is taken from the ARCLake database (MacCallum & Merchant, 2011) and merged with the water composition products.

The potential of the Diversity II inland water database is demonstrated in the context of several case studies, wherein local experts assess the information content against the background of recent events and known environmental trends in individual lakes, including Lakes Biwa, Geneva, Nicaragua, Turkana, Vanern, Victoria and Winnebago. Evidence is presented for eutrophication and hypoxia events, floating vegetation proliferation and the occurrence of cyanobacteria blooms, and reference data from various sources is provided for verification and interpretation of these phenomena. The database is accompanied by a collection of Python functions for generic display tasks and analyses. The functions allow for plotting individual maps for each parameter included in the level 3 products, they perform the extraction of statistical quantities for user-defined regions of interest and display the extracted quantities as time series.

The work done in DIVERSITY II is pioneering work towards atmospheric correction and in-water algorithms globally applied for lakes and water reservoirs, as well as with respect to contrast indicators derived from remote sensing with biodiversity indicators, such as species counts. The results of the project show current possibilities, future improvements by Sentinel 3 as well as by Sentinel 2, and gives recommendations for future work to better link biodiversity in inland water with space observations.