of 48
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.

REPORT OCEANOGRAPHY No. 61, 2017 Linking process rates with modelling data and ecosystem characteristics

Category:

Travel

Publish on:

Views: 3 | Pages: 48

Extension: PDF | Download: 0

Share
Description
REPORT OCEANOGRAPHY No. 61, 2017 Linking process rates with modelling data and ecosystem characteristics Kari Eilola, Stina Lindqvist, Elin Almroth-Rosell, Moa Edman, Iréne Wåhlström, Marco Bartoli, Dorota
Transcript
REPORT OCEANOGRAPHY No. 61, 2017 Linking process rates with modelling data and ecosystem characteristics Kari Eilola, Stina Lindqvist, Elin Almroth-Rosell, Moa Edman, Iréne Wåhlström, Marco Bartoli, Dorota Burska, Jacob Carstensen, Dana Hellemann, Susanna Hietanen, Stefan Hulth, Urszula Janas, Halina Kendzierska, Dorota Pryputniewicz-Flis, Maren Voss, Mindaugas Zilius Front cover: View on Bornö institute for ocean and climate studies. Bornö Station is a marine research station on the island of Stora Bornö in Gullmar Fjord, Bohuslän, on the Swedish West Coast. The complex was built in 1902 by Gustaf Ekman and Otto Pettersson and has been called the cradle of Swedish marine research ( ISSN Linking process rates with modelling data and ecosystem characteristics Authors: Kari Eilola 1, Stina Lindqvist 2, Elin Almroth-Rosell 1, Moa Edman 1, Iréne Wåhlström 1,Marco Bartoli 3, Dorota Burska 4, Jacob Carstensen 5, Dana Hellemann 6, Susanna Hietanen 6, Stefan Hulth 2, Urszula Janas 4, Halina Kendzierska 4, Dorota Pryputniewicz-Flis 4, Maren Voss 7, Mindaugas Zilius 3 1 Swedish Meteorological and Hydrological Institute, Norrköping, Sweden 2 Department of Chemistry and Molecular Biology, University of Gothenburg, Gothenburg, Sweden 3 Klaipeda University, Lithuania 4 Institute of Oceanography, University of Gdansk, Poland 5 Aarhus University, Denmark 6 Department of Environmental Sciences, University of Helsinki, Helsinki, Finland 7 Leibniz Institute for Baltic Sea Research Warnemünde, Germany Corresponding author Kari Eilola, Swedish Meteorological and Hydrological Institute, Sven Källfelts gata 15, SE V Frölunda, Sweden. Tel: +46(0) Summary This report is related to the BONUS project Nutrient Cocktails in COAstal zones of the Baltic Sea alias COCOA. The aim of BONUS COCOA is to investigate physical, biogeochemical and biological processes in a combined and coordinated fashion to improve the understanding of the interaction of these processes on the removal of nutrients along the land-sea interface. The report is especially related to BONUS COCOA WP 6 in which the main objective is extrapolation of results from the BONUS COCOA learning sites to coastal sites around the Baltic Sea in general. Specific objectives of this deliverable (D6.4) were to connect observed process rates with modelling data and ecosystem characteristics. In the report we made statistical analyses of observations from BONUS COCOA study sites together with results from the Swedish Coastal zone Model (SCM). Eight structural variables (water depth, temperature, salinity, bottom water concentrations of oxygen, ammonium, nitrate and phosphate, as well as nitrogen content in sediment) were found common to both the experimentally determined and the model data sets. The observed process rate evaluated in this report was denitrification. In addition regressions were tested between observed denitrification rates and several structural variables (latitude, longitude, depth, light, temperature, salinity, grain class, porosity, loss of ignition, sediment organic carbon, total nitrogen content in the sediment, sediment carbon/nitrogen-ratio, sediment chlorphyll-a as well as bottom water concentrations of oxygen, ammonium, nitrate, and dissolved inorganic phosphorus and silicate) for pooled data from all learning sites. The statistical results showed that experimentally determined multivariate data set from the shallow, illuminated stations was mainly found to be similar to the multivariate data set produced by the SCM model. Generally, no strong correlations of simple relations between observed denitrification and available structural variables were found for data collected from all the learning sites. We found some non-significant correlation between denitrification rates and bottom water dissolved inorganic phosphorous and dissolved silica but the reason behind the correlations is not clear. We also developed and evaluated a theory to relate process rates to monitoring data and nutrient retention. The theoretical analysis included nutrient retention due to denitrification as well as burial of phosphorus and nitrogen. The theory of nutrient retention showed good correlations with model results. It was found that area-specific nitrogen and phosphorus retention capacity in a sub-basin depend much on mean water depth, water residence time, basin area and the mean nutrient concentrations in the active sediment layer and in the water column. Keywords: Coastal zone, Eutrophication, Biogeochemistry, Nutrient retention. 4 Contents 1. Introduction Methods The model data The statistical approach Theoretical modeling of nutrient retention Results and discussion Statistical analysis Description of data Data analysis Theory Importance of physical characteristics on nutrient retention Biogeochemical impact on nutrient retention Regional differences in nutrient retention Theory related to monitoring Other evaluations not shown The benthic primary production Conclusions Acknowledgement References Appendix Some model characteristics 1. Introduction The work in this report, Deliverable (D) 6.4 is related to work package (WP) 6 in the BONUS project Nutrient Cocktails in COAstal zones of the Baltic Sea alias COCOA. The aim within the BONUS COCOA is to investigate physical, biogeochemical and biological processes in a combined and coordinated fashion to improve the understanding of the interaction of these processes on the removal of nutrients along the land-sea interface. The results from the project will be used to estimate nutrient retention capacity in the coastal zone of the entire Baltic Sea coast. Retention studies from the project have been published e.g. by Almroth et al. (2016) who performed a modelling study of nutrient retention in the Stockholm archipelago and by Asmala et al. (2017) who compiled removal rates from coastal systems around the Baltic Sea and analyzed their spatial variation and regulating environmental factors. In BONUS COCOA WP 6, the main objective is extrapolation of results from the learning sites to coastal sites around the Baltic Sea in general. Specific objectives of D6.4 are to connect observed process rates with modelling data and ecosystem characteristics. The aim is to combine information from observed process rates from WP2 to WP4 and from modelling efforts in WP5 to investigate if there are any statistical relations that link process rates with general characteristics obtained from analyses of monitoring data. Most measurements made available for this study have been performed in shallow areas including the illuminated bottoms. A few observed process rates (denitrification) were also available from some deeper non-illuminated locations, the Öre and Vistula estuaries as well as in the Tvärminne archipelago. No monitoring data from the study sites were available for the analysis at the time for the field observations. We therefore focused on the examination of relations between modeled state variables and modeled rates. The model data that have been used for the statistical analysis are extracted results from the Swedish Coastal zone Model (SCM) that cover the entire Swedish Coastal area with 653 sub-basins as described in the BONUS COCOA project D5.1 published as an oceanographic report at SMHI (Eilola et al. 2015). Almroth-Rosell et al. (2016) used the SCM model to discuss modelling of nutrient retention in the Stockholm archipelago. Eight structural variables were found common to both the experimentally determined and the model data sets. The observed process rate evaluated in this report was denitrification. The statistical approach utilizes Principal Components Analysis (PCA) to investigate resemblance in characteristics and denitrification between observed data from the learning sites and modelled sub-basins. In addition regressions were tested between observed denitrification rates and several structural variables. As a complement to the statistical efforts a theoretical analysis was evaluated against the results from the SCM model. The aim was to further investigate and describe the potential links between process rates and nutrient retention to ecosystem characteristics and monitoring data. The analysis included nutrient retention due to denitrification as well as burial of phosphorus (P) and nitrogen (N). Retention of nutrients supplied to a sub-basin can be temporal or permanent as discussed in more detail by Almroth-Rosell et al. (2016). Permanent retention removes the supplied nutrients permanently from the coupled benthic-pelagic biogeochemical cycling under the time scales considered. The temporal retention, i.e. changes in the storage of nutrients that are still active in the biogeochemical cycling during a studied period, can be negative or positive depending on changes in the pelagic and benthic inventory of nutrients. Burial is the only retention process that permanently removes P. For N, in addition to burial, benthic and pelagic denitrification is also defined as permanent retention. Nitrogen fixation adds N and can thereby influence the net N-removal. In the present study, 6 we focus on the permanent retention efficiency of nutrient supplies defined as the internal loss divided by the total supplies of N and P from land, air and surrounding seas to each water body. In addition, we also calculate the area-specific permanent retention efficiency. 7 2. Methods 2.1 The model data SMHI has developed a model system called the Swedish Coastal zone Model (SCM) for water quality calculations in the coastal waters around Sweden. The SCM calculates the state of the water bodies along the entire Swedish coast which is divided according to the Swedish water districts into 5 different parts, and one water district is further divided in 4 parts, resulting in 8 evaluated areas (Fig. 1). The names and the number of subbasin in each area are; the Bothnian Bay 113 sub-basins, the Bothnian Sea 85 sub-basins, the northern Baltic Sea 167 sub-basins, the Östergötland coast 47 sub-basins, the Småland coast 55 sub-basins, the Gotland coast 21 sub-basins, the Skåne-Blekinge coast 52 sub-basins, and finally, the West coast has 113 sub-basins. The hydro-dynamical part of the SCM model calculates with high temporal resolution (10 minutes time step for hydrodynamics) changes in the physical characteristics, including e.g. diurnal variations, freshwater and nutrient supplies, water exchanges and transports of substances between the sub-basins. The biogeochemical model coupled to SCM calculates the changes (1 hour time step for the biogeochemistry) caused by biogeochemical sources and sinks in the sub-basins. The model is described in more detail in D5.1 (Eilola et al. 2015) and in the paper by Almroth-Rosell et al. (2016). Figure 1. The eight different water districts of the SCM model domain. In the present study, we used SCM data from the 5-year period for the statistical analysis of model results in comparison with observed process rates. The model output is vertically integrated instantaneous values for each basin that are given at 7 days intervals, plus additional output at the beginning and end of each month. All modeled process rates are summarized during 24 hours. They are thus not representative for day or night time separately. Note also that even when observed data are from a specific depth range, the model 8 data are still vertically integrated over the entire depth of the water body. This affect the direct comparison with observations since modelled rates from the sediments are averages for the entire area of a sub-basin while the observations are made at specific points within sub-basins. For the evaluation of the theoretical results we used average SCM data for the period Also these values were vertically integrated. The different processes that affect retention have been calculated separately, as they are included in the biogeochemical model SCOBI (Almroth-Rosell et al. 2016). In the present report, we calculate permanent retention efficiency, R, from the SCM model as the internal loss divided by the total supplies of nitrogen and phosphorus from land, air and surrounding seas to each water body. In addition, we also calculate the area-specific permanent retention efficiency defined as R/A where A is the area of the sub-basin. The fraction that is retained is presented in % of the supplies. For N, the permanent internal loss is calculated as the sum of burial and net N 2 production (denitrification - N 2 -fixation). The average rate of water renewal can be estimated from the average age of water (AvA) (Engquist et al. 2006). The water residence time is in the present report calculated as the twenty year mean of the vertical mean AvA in each sub-basin. The concept AvA that has been implemented to the SCM model is described in detail by Engquist et al. (2006). According to their description: This variable is reset to zero for water parcels outside the studied domain. The resulting variable (which represents the specific AvA time of the compartments of the actual subdivision of the domain) is increased one time-step unit for each time step the associated water parcel resides in the domain. In addition to aging, the water parcel is also being subjected to passive tracer advection and diffusion. In time, a quasi-steady state between aging (by remaining in the domain) and rejuvenation (by replacing aged water with new water of zero age) will occur. 9 2.2 The statistical approach The main aim of the present statistical approach was to investigate whether characteristics of observational data gathered at the study sites of BONUS COCOA are represented within the ensemble of modelled sub-basins. At the time when this report is produced there is neither monitoring data nor model results available for the times when observations of process rates are performed within the BONUS COCOA project. This is due to the lack of present forcing conditions for the models since none of the models are run in near-real-time operational mode. On the other hand, a very extensive number of numerical experimental sub-basins were made available for the analysis from the SCM model, but these data do not include areas which were covered by the experimental data made available for the report. The potential to study relations between modelled monitoring data and process rates in the large amount of model results is, however, large. We will therefore investigate if the observed environmental characteristics at the study sites fall within the range of the SCM model results. If so, the statistical analysis will support the potential generality of the conclusions we can get from the theoretical study. The data set available from experimental efforts was compared with output from the SCM model in uni- and multivariate analyses. From a larger data set made available from the SCM model eight structural variables were common to both the experimentally determined and the model data sets: water depth, temperature, salinity, bottom water concentrations of oxygen, ammonium, nitrate and phosphate, as well as N content in sediment. The process rate evaluated in this report was denitrification, which was experimentally determined by 15 N- amendments. To investigate potential simple relations regression analyses (linear and quadratic) were performed with the experimental data set between observed denitrification rates and simultaneous observations of different structural variables (latitude, longitude, depth, light, temperature, salinity, grain class, porosity, loss of ignition (LOI), sediment organic carbon, total nitrogen content in the sediment, sediment carbon/nitrogen ratio (C/Nratio), sediment chlorophyll a (Chl-a) as well as bottom water concentrations of oxygen (O 2 ), ammonium (NH 4 ), nitrate (NO 3 ), and dissolved inorganic phosphorus (DIP) and silicate (DSi) for pooled data from all learning sites. PCAs were performed in Simca (v. 14.0, Umetrics AB, Sweden) to compare observed data with model data in terms of structural variables and denitrification. Where appropriate, variables were log transformed. The variables were mean centered and scaled to unitvariance. The models were diagnosed by R2 and Q2, measuring model fit (R2) and percent of the variation of the data that can be predicted by the model as calculated by cross validation (Q2) (Eriksson et al., 2006). Initial analyses included all eight structural variables common to both data sets. Subsequent analyses were run on a subset of structural variables. Although there were missing data in the experimentally determined variables, five study sites were considered to include sufficient information to run the analyses; Curonian Lagoon, Roskilde Fjord, Puck Bay, Tvärminne and Vistula Estuary (Table 1). Stations included from the Curonian Lagoon were sampled during all seasons in 2014 and From Roskilde Fjord three stations sampled in autumn 2015 were included. Data from two Puck Bay stations were collected during three seasons ( ). At the Tvärminne study site, two illuminated stations were sampled in spring, summer and autumn and two deeper, nonilluminated stations were sampled monthly from spring to autumn in 2015 and Vistula stations were sampled during three seasons in (Table 1). The shallower, illuminated stations and the deeper, non-illuminated stations were analyzed separately as initial analysis indicated different characteristics for observed data. The illuminated stations were evaluated together with model sub-basins with a maximum depth of 10 10 m, which included 129 sub-basins in total. The 24 and 33 m Tvärminne and Vistula Estuary stations were modelled with 121 sub-basins with a maximum depth of 23 to 34 m. PCA was run on all sub-basins within the depth range along the Swedish coast as well as on each of the five Swedish Water Districts (as defined by the County Administrative Boards) for a more detailed analysis. 11 Table 1. Name of the stations and their depth, light condition and time period for the field work for the observation data used in the PCA and for the correlation tests. The data set was not complete from all study sites and the missing variables are indicated in the table. Data from stations not used in the PCA, but used in the correlation tests are shown in the end of the table (Vistula Estuary and Öre Estuary Corr.tests). Study site Station Depth Illum- Year Seasons Missing variables (m) inated Curonian Litoral 1 Yes All Sal(partly), Denitr Lagoon Vidmares 2 Yes All Nida 4 Yes All Roskilde Station 60 5 Yes 2015 Autumn Temp, Sal, Ntot Fjord Zostera 2 Yes 2015 Autumn Sand 2 Yes 2015 Autumn Puck Bay Sand 2 Yes 2015, 2016 Spring/summer/ autumn Sand 5 Yes 2015, 2016 Spring/summer/ autumn Tvärminne Mud 2 Yes 2015, 2016 Spring/summer/ autumn Sal,Denitr, Ntot(partly) Sand 3 Yes 2015, 2016 Spring/summer/ autumn Storfjärden 24 No 2015, 2016 Monthly, springautumn Ntot i30 33 No 2015, 2016 Monthly, spring- Vistula Estuary autumn VE02 20 No 2015, 2016 Winter/spring Ntot, NH4(partly), Denitr VE05 22 No 2014, 2015, Summer/winter/ Ntot, Denitr(partly) 2016 spring VE10 22 No 2015 Winter Ntot, Denitr VE18 24 No 2014, 2016 Summer/spring Ntot, NH4(partly), Denitr(partly) VE49a 26 No 2014, 2016 Summer /spring Ntot VE13 28 No 2014, 2015, Summer/winter/ Ntot, Denitr(partly) 2016 spring VE09 29 No 2014, 2015, Summer/winter/ Ntot 2016 spring VE06 34 No 2015, 2016 Winter/spring Ntot, Denitr(partly) Vistula VE04 13 No 2016 Spring Estuary VE03 16 No 2014 Summer Corr.tests VE15 43 No 2014, 2016 Summer /spring VE46 48 No 2014 Summer VE07 51 No 2016 Spring VE38 68 No 2014, 2016 Summer/spring VE43 94 No 2014 Summer TF No 2014 Summer Öre N3 17 No 2015 Spring Estuary N5 17 No 2015 Summer Corr.tests N34 17 No 2015 Summer N6 18 No 2015 Spring N7 19 No 2015 Summer/Spring N8 19 No 2015 Spring N10 21 No 2015 Summer/Spring N11 24 No 2015 Summer N14 37 No 2015 Summer/Spring NB8 35 No 2015 Summer/Spring 12 2.3 Theoretical modeling of nutrient
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks