Sylwia Szporak-Wasilewska

Developing an algorithm for enhancement of a digital terrain model for a densely vegetated floodplain wetland

Abstract. Airborne laser scanning survey data were conducted with a scanning density of 4  points/m2 to accurately map the surface of a unique central European complex of wetlands: the lower Biebrza River valley (Poland). A method to correct a degrading effect of vegetation (so-called “vegetation effect”) on digital terrain models (DTMs) was applied utilizing remotely sensed images, real-time kinematic global positioning system elevation measurements, topographical surveys, and vegetation height measurements. Geographic object-based image analysis (GEOBIA) was performed to map vegetation within the study area that was used as categories from which vegetation height information was derived for the DTM correction. The final DTM was compared with a model obtained, where additional correction of the “vegetation effect” was neglected. A comparison between corrected and uncorrected DTMs demonstrated the importance of accurate topography through a simple presentation of the discrepancies arising in features of the flood using various DTM products. An overall map classification accuracy of 80% was attained with the use of GEOBIA. Correction factors developed for various types of the vegetation reached values from 0.08 up to 0.92 m and were dependent on the vegetation type.

Processing of airborne laser scanning data to generate accurate DTM for floodplain wetland

Structure of the floodplain, especially its topography and vegetation, influences the overland flow and dynamics of floods which are key factors shaping ecosystems in surface water-fed wetlands. Therefore elaboration of the digital terrain model (DTM) of a high spatial accuracy is crucial in hydrodynamic flow modelling in river valleys. In this study the research was conducted in the unique Central European complex of fens and marshes - the Lower Biebrza river valley. The area is represented mainly by peat ecosystems which according to EU Water Framework Directive (WFD) are called “water-dependent ecosystems”. Development of accurate DTM in these areas which are overgrown by dense wetland vegetation consisting of alder forest, willow shrubs, reed, sedges and grass is very difficult, therefore to represent terrain in high accuracy the airborne laser scanning data (ALS) with scanning density of 4 points/m2 was used and the correction of the “vegetation effect” on DTM was executed. This correction was performed utilizing remotely sensed images, topographical survey using the Real Time Kinematic positioning and vegetation height measurements. In order to classify different types of vegetation within research area the object based image analysis (OBIA) was used. OBIA allowed partitioning remotely sensed imagery into meaningful image-objects, and assessing their characteristics through spatial and spectral scale. The final maps of vegetation patches that include attributes of vegetation height and vegetation spectral properties, utilized both the laser scanning data and the vegetation indices developed on the basis of airborne and satellite imagery. This data was used in process of segmentation, attribution and classification. Several different vegetation indices were tested to distinguish different types of vegetation in wetland area. The OBIA classification allowed correction of the “vegetation effect” on DTM. The final digital terrain model was compared and examined within distinguished land cover classes (formed mainly by natural vegetation of the river valley) with archival height models developed through interpolation of ground points measured with GPS RTK and also with elevation models from the ASTER-GDEM and SRTM programs. The research presented in this paper allowed improving quality of hydrodynamic modelling in the surface water-fed wetlands protected within Biebrza National Park. Additionally, the comparison with other digital terrain models allowed to demonstrate the importance of accurate topography products in such modelling. The ALS data also significantly improved the accuracy and actuality of the river Biebrza course, its tributaries and location of numerous oxbows typical in this part of the river valley in comparison to previously available data. This type of data also helped to refine the river valley cross-sections, designate river banks and to develop the slope map of the research area.

Spatial and temporal variability of the interception in the natural wetland valley, the lower Biebrza basin case study

Abstract Spatial and temporal variability of the interception in the natural wetland valley, the lower Biebrza basin case study. The paper presents the research carried out in the lower basin of Biebrza River valley in order to identify interception for natural wetland plant communities. Maximum interception, i.e. the largest amount of water, expressed in millimeters, which can be captured and retained by plant canopy from rainfall is one of the key parameters of the water cycle modeling. Maximum interception was determined based on the difference of the masses of wet and dry fresh plant samples. Collection of plant material samples took place during the five measurement sessions, which began immediately after the flood recedes, and then lasted until the end of the growing season. Interception spatial variability was analyzed on the basis of the results of maximum interception measured for selected plant aggregations in the different sampling points. The obtained values were extrapolated to the area of the lower basin of Biebrza River using vegetation map of the Biebrza National Park. By conducting a test sessions in the five coming months, the maps of the spatial variability also show changes over time. Methodology used in the described tests allowed for obtaining of satisfactory results. They present, in a correct way, variation occurring between the plant aggregations due to their morphology. In most cases the results are consistent with data from the literature. As results of the analysis of spatial variability of the maximum interception, the highest values were found for the plant communities located in the immediate vicinity of the river channel. With the increase of the distance from river towards the valley edges the maximum interception values decrease. These changes can be seen in the form of strips parallel to the river channel, which corresponds to the plant zones. Obtained map of spatial variability of the maximum interception, which is the results of extrapolation of the values assigned to plant communities, has a high correlation with the map resulting from the analysis of satellite images Streszczenie Przestrzenna i czasowa zmienność intercepcji w naturalnej dolinie bagiennej, analiza przypadku dolnego basenu rzeki Biebrzy. W artykule przedstawiono badania przeprowadzone w dolnym basenie doliny rzeki Biebrzy mające na celu rozpoznanie procesów intercepcji w naturalnych zbiorowiskach roślinnych obszarów mokradłowych. Intercepcja maksymalna, czyli ilość wody wyrażona w milimetrach jaką roślina może przechwycić i zatrzymać z opadu atmosferycznego, jest jednym z kluczowych parametrów modelowania cyklu wodnego. Intercepcja maksymalna została określona na podstawie różnicy mas próbek suchych i mokrych świeżych roślin. Pobór materiału roślinnego odbywał się podczas pięciu sesji pomiarowych, które rozpoczęto bezpośrednio po ustąpieniu zalewu i które trwały do końca okresu wegetacyjnego. Przestrzenna zmienność intercepcji była analizowana na podstawie wyników pomiarów intercepcji maksymalnej dla wybranych zespołów roślinnych w różnych punktach pomiarowych. Otrzymane wartości ekstrapolowano na obszar dolnego basenu Biebrzy z wykorzystaniem mapy roślinności Biebrzańskiego Parku Narodowego. Dzięki przeprowadzeniu sesji pomiarowych w kolejnych miesiącach, mapy zmienności przestrzennej przedstawiają również zmiany zachodzące w czasie. Zastosowana w opisanych badaniach metodyka pozwoliła na uzyskanie zadawalających wyników. Prezentują one w sposób prawidłowy zmienność występującą między zespołami roślinnymi, związaną z ich morfologią. W większości przypadków wyniki są zgodne z danymi z literatury. W wyniku analizy przestrzennej zmienności intercepcji maksymalnej najwyższe wartości stwierdzono dla zbiorowisk położonych w bezpośrednim sąsiedztwie koryta rzecznego. Wraz ze wzrostem odległości od rzeki ku brzegom doliny wartości intercepcji maksymalnej maleją. Zmiany widoczne są w postaci pasów równoległych do koryta rzecznego, które odpowiadają strefom roślinnym. Mapa zmienności przestrzennej intercepcji maksymalnej, która powstała w wyniku ekstrapolacji wartości przypisanych zespołom roślinnym, wykazuje dużą korelację z mapą powstałą w wyniku analizy zdjęcia satelitarnego.

Too wet and too dry? Uncertainty of DEM as a potential source of significant errors in a model-based water level assessment in riparian and mire ecosystems

Modelling groundwater depths in floodplains and peatlands remains a basic approach to assessing hydrological conditions of habitats. Groundwater flow models used to compute groundwater heads are known for their uncertainties, and the calibration of these models and the uncertainty assessments of parameters remain fundamental steps in providing reliable data. However, the elevation data used to determine the geometry of model domains are frequently considered deterministic and hence are seldom considered a source of uncertainty in model-based groundwater level estimations. Knowing that even the cutting-edge laser-scanning-based digital elevation models have errors due to vegetation effects and scanning procedure failures, we provide an assessment of uncertainty of water level estimations that remain basic data for wetland ecosystem assessment and management. We found that the uncertainty of the digital elevation model (DEM) significantly influenced the results of the assessment of the habitat’s hydrological conditions expressed as groundwater depths. In extreme cases, although the average habitat suitability index (HSI) assessed in a deterministic manner was defined as ‘unsuitable’, in a probabilistic approach (grid-cell-scale estimation), it reached a value of 40% probability, signifying ‘optimum’ or ‘tolerant’. For the 24 habitats analysed, we revealed vast differences between HSI scores calculated for individual grid cells of the model and HSI scores computed as average values from the set of grid cells located within the habitat patches. We conclude that groundwater-modelling-based decision support approaches to wetland assessment can result in incorrect management if the quality of DEM has not been addressed in studies referring to groundwater depths.

Towards rainfall interception capacity estimation using ALS LiDAR data

In this study we develop a spatial model for interception capacity of vegetation based on LiDAR data. The study is conducted in the natural wetland river valley dominated meadows, reeds and small bushes. The multiple regression model was chosen to relate the field measurements of interception capacity and LiDAR statistics at 2m grid. The optimal model was chosen by stepwise selection and further manual variables selection resulting in the r2 of 0.52 and the residual standard error of 0.27 mm. The model preserved the vegetation pattern spatially and showed reasonable estimates for both vegetation covered and not covered by field sampling. The model was, however, affected by LiDAR measurements corrupted by river inundation. The results show good perspective for using LiDAR data for interception capacity estimation.

Derivation from the Landsat 7 NDVI and ground truth validation of LAI and interception storage capacity for wetland ecosystems in Biebrza Valley, Poland

Wetlands are very valuable areas because they provide a wide range of ecosystems services therefore modeling of wetland areas is very relevant, however, the most widely used hydrological models were developed in the 90s and usually are not adjusted to simulate wetland conditions. In case of wetlands including interception storage into the model’s calculation is even more challenging, because literature data hardly exists. This study includes the computation of interception storage capacity based on Landsat 7 image and ground truthing measurements conducted in the Biebrza Valley, Poland. The method was based on collecting and weighing dry, wet and fully saturated samples of sedges. During the experiments measurements of fresh/dry biomass and leaf area index (LAI) were performed. The research was repeated three times during the same season (May, June and July 2013) to observe temporal variability of parameters. Ground truthing measurements were used for the validating estimation of parameters derived from images acquired in a similar period as the measurements campaigns. The use of remote sensing has as major advantage of being able to obtain an area covering spatially and temporally distributed estimate of the interception storage capacity. Results from this study proved that interception capacity of wetlands vegetation is changing considerably during the vegetation season (temporal variability) and reaches its maximum value when plants are fully developed. Different areas depending on existing plants species are characterized with different values of interception capacity (spatial variability). This research frames within the INTREV and HiWET projects, funded respectively by National Science Centre (NCN) in Poland and BELSPO STEREO III.