Retrieval of grassland biophysical parameters using multitemporal optical and radar satellite data

The amount and quality of available biomass is a key factor for the sustainable livestock industry and agricultural management related decision making. Globally 31.5% of land cover is grassland while 80% of Ireland’s agricultural land is grassland. In Ireland, grasslands are intensively managed and provide the cheapest feed source for animals. This dissertation presents a detailed state of the art review of satellite remote sensing of grasslands, and the potential application of optical (Moderate–resolution Imaging Spectroradiometer (MODIS)) and radar (TerraSAR-X) time series imagery to estimate the grassland biomass at two study sites (Moorepark and Grange) in the Republic of Ireland using both statistical and state of the art machine learning algorithms. High quality weather data available from the on-site weather station was also used to calculate the Growing Degree Days (GDD) for Grange to determine the impact of ancillary data on biomass estimation. In situ and satellite data covering 12 years for the Moorepark and 6 years for the Grange study sites were used to predict grassland biomass using multiple linear regression, Neuro Fuzzy Inference Systems (ANFIS) models. The results demonstrate that a dense (8-day composite) MODIS image time series, along with high quality in situ data, can be used to retrieve grassland biomass with high performance (R2 = 0:86; p < 0:05, RMSE = 11.07 for Moorepark). The model for Grange was modified to evaluate the synergistic use of vegetation indices derived from remote sensing time series and accumulated GDD information. As GDD is strongly linked to the plant development, or phonological stage, an improvement in biomass estimation would be expected. It was observed that using the ANFIS model the biomass estimation accuracy increased from R2 = 0:76 (p < 0:05) to R2 = 0:81 (p < 0:05) and the root mean square error was reduced by 2.72%. The work on the application of optical remote sensing was further developed using a TerraSAR-X Staring Spotlight mode time series over the Moorepark study site to explore the extent to which very high resolution Synthetic Aperture Radar (SAR) data of interferometrically coherent paddocks can be exploited to retrieve grassland biophysical parameters. After filtering out the non-coherent plots it is demonstrated that interferometric coherence can be used to retrieve grassland biophysical parameters (i. e., height, biomass), and that it is possible to detect changes due to the grass growth, and grazing and mowing events, when the temporal baseline is short (11 days). However, it not possible to automatically uniquely identify the cause of these changes based only on the SAR backscatter and coherence, due to the ambiguity caused by tall grass laid down due to the wind. Overall, the work presented in this dissertation has demonstrated the potential of dense remote sensing and weather data time series to predict grassland biomass using machine-learning algorithms, where high quality ground data were used for training. At present a major limitation for national scale biomass retrieval is the lack of spatial and temporal ground samples, which can be partially resolved by minor modifications in the existing PastureBaseIreland database by adding the location and extent ofeach grassland paddock in the database. As far as remote sensing data requirements are concerned, MODIS is useful for large scale evaluation but due to its coarse resolution it is not possible to detect the variations within the fields and between the fields at the farm scale. However, this issue will be resolved in terms of spatial resolution by the Sentinel-2 mission, and when both satellites (Sentinel-2A and Sentinel-2B) are operational the revisit time will reduce to 5 days, which together with Landsat-8, should enable sufficient cloud-free data for operational biomass estimation at a national scale. The Synthetic Aperture Radar Interferometry (InSAR) approach is feasible if there are enough coherent interferometric pairs available, however this is difficult to achieve due to the temporal decorrelation of the signal. For repeat-pass InSAR over a vegetated area even an 11 days temporal baseline is too large. In order to achieve better coherence a very high resolution is required at the cost of spatial coverage, which limits its scope for use in an operational context at a national scale. Future InSAR missions with pair acquisition in Tandem mode will minimize the temporal decorrelation over vegetation areas for more focused studies. The proposed approach complements the current paradigm of Big Data in Earth Observation, and illustrates the feasibility of integrating data from multiple sources. In future, this framework can be used to build an operational decision support system for retrieval of grassland biophysical parameters based on data from long term planned optical missions (e. g., Landsat, Sentinel) that will ensure the continuity of data acquisition. Similarly, Spanish X-band PAZ and TerraSAR-X2 missions will ensure the continuity of TerraSAR-X and COSMO-SkyMed.

Collection of aboveground community and species-specific plant biomass from the Jena Experiment (time series since 2002).

This data set comprises time series of aboveground community plant biomass (Sown plant community, Weed plant community, Dead plant material, and Unidentified plant material; all measured in biomass as dry weight) and species-specific biomass from the sown species of several experiments at the field site of a large grassland biodiversity experiment (the Jena Experiment; see further details below). Aboveground community biomass was normally harvested twice a year just prior to mowing (during peak standing biomass twice a year, generally in May and August; in 2002 only once in September) on all experimental plots in the Jena Experiment. This was done by clipping the vegetation at 3 cm above ground in up to four rectangles of 0.2 x 0.5 m per large plot. The location of these rectangles was assigned by random selection of new coordinates every year within the core area of the plots. The positions of the rectangles within plots were identical for all plots. The harvested biomass was sorted into categories: individual species for the sown plant species, weed plant species (species not sown at the particular plot), detached dead plant material (i.e., dead plant material in the data file), and remaining plant material that could not be assigned to any category (i.e., unidentified plant material in the data file). All biomass was dried to constant weight (70°C, >= 48 h) and weighed. Sown plant community biomass was calculated as the sum of the biomass of the individual sown species. The data for individual samples and the mean over samples for the biomass measures on the community level are given. Overall, analyses of the community biomass data have identified species richness as well as functional group composition as important drivers of a positive biodiversity-productivity relationship. The following series of datasets are contained in this collection: 1. Plant biomass form the Main Experiment: In the Main Experiment, 82 grassland plots of 20 x 20 m were established from a pool of 60 species belonging to four functional groups (grasses, legumes, tall and small herbs). In May 2002, varying numbers of plant species from this species pool were sown into the plots to create a gradient of plant species richness (1, 2, 4, 8, 16 and 60 species) and functional richness (1, 2, 3, 4 functional groups). 2. Plant biomass from the Dominance Experiment: In the Dominance Experiment, 206 grassland plots of 3.5 x 3.5 m were established from a pool of 9 species that can be dominant in semi-natural grassland communities of the study region. In May 2002, varying numbers of plant species from this species pool were sown into the plots to create a gradient of plant species richness (1, 2, 3, 4, 6, and 9 species). 3. Plant biomass from the monoculture plots: In the monoculture plots the sown plant community contains only a single species per plot and this species is a different one for each plot. Which species has been sown in which plot is stated in the plot information table for monocultures (see further details below). The monoculture plots of 3.5 x 3.5 m were established for all of the 60 plant species of the Jena Experiment species pool with two replicates per species like the other experiments in May 2002. All plots were maintained by bi-annual weeding and mowing.

Collection of data on aboveground invertebrates in the Jena Experiment (time series since 2002)

This collection contains measurements of abundance and diversity of different groups of aboveground invertebrates sampled on the plots of the different sub-experiments at the field site of a large grassland biodiversity experiment (the Jena Experiment; see further details below). In the main experiment, 82 grassland plots of 20 x 20 m were established from a pool of 60 species belonging to four functional groups (grasses, legumes, tall and small herbs). In May 2002, varying numbers of plant species from this species pool were sown into the plots to create a gradient of plant species richness (1, 2, 4, 8, 16 and 60 species) and functional richness (1, 2, 3, 4 functional groups). Plots were maintained by bi-annual weeding and mowing. The following series of datasets are contained in this collection: 1. Measurements of ant abundance (number of individuals attracted to baits) and ant occurrence (binary data) in the Main Experiment in 2006 and 2013. Ants where sampled using two types of baited traps receiving ~10g of Tuna or ~10g of honey/Sucrose. After 30min the occurrence (presence = 1 / absence = 0) and abundance (number) of ants at the two types of baits was recorded and pooled per plot.

Collection of vegetation and soil surface cover in the Jena Experiment (time series since 2002)

This collection contains measurements of vegetation and soil surface cover measured on the plots of the different sub-experiments at the field site of a large grassland biodiversity experiment (the Jena Experiment; see further details below). In the main experiment, 82 grassland plots of 20 x 20 m were established from a pool of 60 species belonging to four functional groups (grasses, legumes, tall and small herbs). In May 2002, varying numbers of plant species from this species pool were sown into the plots to create a gradient of plant species richness (1, 2, 4, 8, 16 and 60 species) and functional richness (1, 2, 3, 4 functional groups). Plots were maintained by bi-annual weeding and mowing. The following series of datasets are contained in this collection: 1. Measurements of vegetation cover, i.e. the proportion of soil surface area that is covered by different categories of plants per estimated plot area. Data was collected on the plant community level (sown plant community, weed plant community, dead plant material, and bare ground) and on the level of individual plant species in case of the species that have been sown into the plots to create the gradient of plant diversity.

Collection of measurements on physical soil properties in plots of the Jena Experiment (time series since 2002)

This collection contains measurements on physical soil properties of the plots of the different sub-experiments at the field site of a large grassland biodiversity experiment (the Jena Experiment; see further details below). In the main experiment, 82 grassland plots of 20 x 20 m were established from a pool of 60 species belonging to four functional groups (grasses, legumes, tall and small herbs). In May 2002, varying numbers of plant species from this species pool were sown into the plots to create a gradient of plant species richness (1, 2, 4, 8, 16 and 60 species) and functional richness (1, 2, 3, 4 functional groups). Plots were maintained by bi-annual weeding and mowing

Aboveground community and species-specific plant biomass from the Jena Experiment (Dominance Experiment, year 2008)

This data set contains aboveground community plant biomass (Sown plant community, Weed plant community, Dead plant material, and Unidentified plant material; all measured in biomass as dry weight) and species-specific biomass from the sown species of the dominance experiment plots of a large grassland biodiversity experiment (the Jena Experiment; see further details below). In the dominance experiment, 206 grassland plots of 3.5 x 3.5 m were established from a pool of 9 plant species that can be dominant in semi-natural grassland communities of the study region. In May 2002, varying numbers of plant species from this species pool were sown into the plots to create a gradient of plant species richness (1, 2, 3, 4, 6, and 9 species). Plots were maintained by bi-annual weeding and mowing. Aboveground community biomass was harvested twice in May and August 2008 on all experimental plots of the dominance experiment. This was done by clipping the vegetation at 3 cm above ground in two rectangles of 0.2 x 0.5 m per experimental plot. The location of these rectangles was assigned by random selection of coordinates within the central area of the plots (excluding an outer edge of 50cm). The positions of the rectangles within plots were identical for all plots. The harvested biomass was sorted into categories: individual species for the sown plant species, weed plant species (species not sown at the particular plot), detached dead plant material, and remaining plant material that could not be assigned to any category. All biomass was dried to constant weight (70°C, >= 48 h) and weighed. Sown plant community biomass was calculated as the sum of the biomass of the individual sown species. The mean of both samples per plot and the individual measurements are provided in the data file. Overall, analyses of the community biomass data have identified species richness and the presence of particular species as an important driver of a positive biodiversity-productivity relationship.

Aboveground plant community and species-specific vegetation cover from the Jena Experiment (Main Experiment, year 2009)

This data set contains information on vegetation cover, i.e. the proportion of soil surface area that is covered by different categories of plants per estimated plot area. Data was collected on the plant community level (sown plant community, weed plant community, dead plant material, and bare ground) and on the level of individual plant species in case of the sown species. Data presented here is from the Main Experiment plots of a large grassland biodiversity experiment (the Jena Experiment; see further details below). In the main experiment, 82 grassland plots of 20 x 20 m were established from a pool of 60 species belonging to four functional groups (grasses, legumes, tall and small herbs). In May 2002, varying numbers of plant species from this species pool were sown into the plots to create a gradient of plant species richness (1, 2, 4, 8, 16 and 60 species) and functional richness (1, 2, 3, 4 functional groups). Plots were maintained by bi-annual weeding and mowing. In 2009, vegetation cover was estimated twice in May and August just prior to mowing (during peak standing biomass) on all experimental plots of the Main Experiment. Cover was visually estimated in a central area of each plot 3 by 3 m in size (approximately 9 m²) using a decimal scale (Londo). Cover estimates for the individual species (and for target species + weeds + bare ground) can add up to more than 100% because the estimated categories represented a structure with potentially overlapping multiple layers. In 2009, in addition to the four community level cover estimates, cover of the moss layer was estimated.

Aboveground plant community and species-specific vegetation cover from the Jena Experiment (Main Experiment, year 2010)

This data set contains information on vegetation cover, i.e. the proportion of soil surface area that is covered by different categories of plants per estimated plot area. Data was collected on the plant community level (sown plant community, weed plant community, dead plant material, and bare ground) and on the level of individual plant species in case of the sown species. Data presented here is from the Main Experiment plots of a large grassland biodiversity experiment (the Jena Experiment; see further details below). In the main experiment, 82 grassland plots of 20 x 20 m were established from a pool of 60 species belonging to four functional groups (grasses, legumes, tall and small herbs). In May 2002, varying numbers of plant species from this species pool were sown into the plots to create a gradient of plant species richness (1, 2, 4, 8, 16 and 60 species) and functional richness (1, 2, 3, 4 functional groups). Plots were maintained by bi-annual weeding and mowing. In 2010, vegetation cover was estimated twice in May and August just prior to mowing (during peak standing biomass) on all experimental plots of the Main Experiment. Cover was visually estimated in a central area of each plot 3 by 3 m in size (approximately 9 m²) using a decimal scale (Londo). Cover estimates for the individual species (and for target species + weeds + bare ground) can add up to more than 100% because the estimated categories represented a structure with potentially overlapping multiple layers.

Aboveground plant community and species-specific vegetation cover from the Jena Experiment (Main Experiment, year 2013)

This data set contains information on vegetation cover, i.e. the proportion of soil surface area that is covered by different categories of plants per estimated plot area. Data was collected on the plant community level (sown plant community, weed plant community, dead plant material, and bare ground) and on the level of individual plant species in case of the sown species. Data presented here is from the Main Experiment plots of a large grassland biodiversity experiment (the Jena Experiment; see further details below). In the main experiment, 82 grassland plots of 20 x 20 m were established from a pool of 60 species belonging to four functional groups (grasses, legumes, tall and small herbs). In May 2002, varying numbers of plant species from this species pool were sown into the plots to create a gradient of plant species richness (1, 2, 4, 8, 16 and 60 species) and functional richness (1, 2, 3, 4 functional groups). Plots were maintained by bi-annual weeding and mowing. In 2013, vegetation cover was estimated twice in May and August just prior to mowing (during peak standing biomass) on all experimental plots of the Main Experiment. Cover was visually estimated in a central area of each plot 3 by 3 m in size (approximately 9 m²) using a decimal scale (Londo). Cover estimates for the individual species (and for target species + weeds + bare ground) can add up to more than 100% because the estimated categories represented a structure with potentially overlapping multiple layers.

Aboveground plant community and species-specific vegetation cover from the Jena Experiment (Main Experiment, year 2008)

This data set contains information on vegetation cover, i.e. the proportion of soil surface area that is covered by different categories of plants per estimated plot area. Data was collected on the plant community level (sown plant community, weed plant community, dead plant material, and bare ground) and on the level of individual plant species in case of the sown species. Data presented here is from the Main Experiment plots of a large grassland biodiversity experiment (the Jena Experiment; see further details below). In the main experiment, 82 grassland plots of 20 x 20 m were established from a pool of 60 species belonging to four functional groups (grasses, legumes, tall and small herbs). In May 2002, varying numbers of plant species from this species pool were sown into the plots to create a gradient of plant species richness (1, 2, 4, 8, 16 and 60 species) and functional richness (1, 2, 3, 4 functional groups). Plots were maintained by bi-annual weeding and mowing. In 2008, vegetation cover was estimated twice in May and August just prior to mowing (during peak standing biomass) on all experimental plots of the Main Experiment. Cover was visually estimated in a central area of each plot 3 by 3 m in size (approximately 9 m²) using a decimal scale (Londo). Cover estimates for the individual species (and for target species + weeds + bare ground) can add up to more than 100% because the estimated categories represented a structure with potentially overlapping multiple layers.

Soil temperature along a plant diversity gradient in the Jena Experiment (Main Experiment, 2013)

Soil temperature (in °C) was determined using a frequency domain sensor probe (WET-2 Sensor, Delta-T Devices, Cambridge, United Kingdom) on 1st August 2013. The device was inserted from the top 6 cm deep (length of the prongs) into the soil. The average of three measurements on the same day was calculated. All data where measured in the main experiment plots of a large grassland biodiversity experiment (the Jena Experiment; see further details below). In the main experiment, 82 grassland plots of 20 x 20 m were established from a pool of 60 species belonging to four functional groups (grasses, legumes, tall and small herbs). In May 2002, varying numbers of plant species from this species pool were sown in the plots to create a gradient of plant species richness (1, 2, 4, 8, 16 and 60 species) and functional richness (1, 2, 3, or 4 functional groups). Plots were maintained by bi-annual weeding and mowing.

Soil porosity along a plant diversity gradient in the Jena Experiment (Main Experiment, 2013)

Soil porosity is the fraction of total volume occupied by pores or voids measured at matric potential 0. To measure soil porosity, soil samples were taken from each plot using sample rings with an internal diameter of 57 mm and height of 40.5 mm (inner volume of Vs=100 cm3). The samples were placed on a sand bed box with water level set to allow saturation of the samples with water. After 48 h the samples were weighed (ms), oven dried at 105 °C and weighed again to determine the dry weight (md). We calculated soil porosity (n [%]) using the density of water (?w=1 g cm?3), n=100 ? (mw-md) / (?w?Vs). To account for the spatial variation of soil properties, three replicates were taken per plot, approximately 2, 3 and 4 weeks after the flood that occurred at the field site during June 2013. Data are the average soil porosity values per plot. All data where measured in the main experiment plots of a large grassland biodiversity experiment (the Jena Experiment; see further details below). In the main experiment, 82 grassland plots of 20 x 20 m were established from a pool of 60 species belonging to four functional groups (grasses, legumes, tall and small herbs). In May 2002, varying numbers of plant species from this species pool were sown in the plots to create a gradient of plant species richness (1, 2, 4, 8, 16 and 60 species) and functional richness (1, 2, 3, or 4 functional groups). Plots were maintained by bi-annual weeding and mowing.

Aboveground community and species-specific plant biomass from the Jena Experiment (Main Experiment, year 2009)

This data set contains aboveground community biomass in 2009 (Sown plant community, Weed plant community, Dead plant material, and Unidentified plant material; all measured in biomass as dry weight) and species-specific biomass from the sown species of the main experiment plots of a large grassland biodiversity experiment (the Jena Experiment; see further details below). In the main experiment, 82 grassland plots of 20 x 20 m were established from a pool of 60 species belonging to four functional groups (grasses, legumes, tall and small herbs). In May 2002, varying numbers of plant species from this species pool were sown into the plots to create a gradient of plant species richness (1, 2, 4, 8, 16 and 60 species) and functional richness (1, 2, 3, 4 functional groups). Plots were maintained by bi-annual weeding and mowing. Aboveground community biomass was harvested twice in 2009 just prior to mowing (during peak standing biomass in early June and in late August) on all experimental plots of the main experiment. This was done by clipping the vegetation at 3 cm above ground in three rectangles of 0.2 x 0.5 m per large plot. The location of these rectangles was assigned prior to each harvest by random selection of coordinates within the core area of the plots (i.e. the central 10 x 15 m). The positions of the rectangles within plots were identical for all plots. The harvested biomass was sorted into categories: individual species for the sown plant species, weed plant species (species not sown at the particular plot), detached dead plant material (i.e., dead plant material in the data file), and remaining plant material that could not be assigned to any category (i.e., unidentified plant material in the data file). All biomass was dried to constant weight (70°C, >= 48 h) and weighed. Sown plant community biomass was calculated as the sum of the biomass of the individual sown species. The data for individual samples and the mean over samples for all biomass measures are given. Overall, analyses of the community biomass data have identified species richness as well as functional group composition as important drivers of a positive biodiversity-productivity relationship.

Aboveground community and species-specific plant biomass from the Jena Experiment (Main Experiment, year 2010)

This data set contains aboveground community biomass in 2010 (Sown plant community, Weed plant community, Dead plant material, and Unidentified plant material; all measured in biomass as dry weight) and species-specific biomass from the sown species of the main experiment plots of a large grassland biodiversity experiment (the Jena Experiment; see further details below). In the main experiment, 82 grassland plots of 20 x 20 m were established from a pool of 60 species belonging to four functional groups (grasses, legumes, tall and small herbs). In May 2002, varying numbers of plant species from this species pool were sown into the plots to create a gradient of plant species richness (1, 2, 4, 8, 16 and 60 species) and functional richness (1, 2, 3, 4 functional groups). Plots were maintained by bi-annual weeding and mowing. Aboveground community biomass was harvested twice in 2010 just prior to mowing (during peak standing biomass in early June and in late August) on all experimental plots of the main experiment. This was done by clipping the vegetation at 3 cm above ground in two rectangles of 0.2 x 0.5 m per large plot. The location of these rectangles was assigned prior to each harvest by random selection of coordinates within the core area of the plots (i.e. the central 10 x 15 m). The positions of the rectangles within plots were identical for all plots. The harvested biomass was sorted into categories: individual species for the sown plant species, weed plant species (species not sown at the particular plot), detached dead plant material (i.e., dead plant material in the data file), and remaining plant material that could not be assigned to any category (i.e., unidentified plant material in the data file). All biomass was dried to constant weight (70°C, >= 48 h) and weighed. Sown plant community biomass was calculated as the sum of the biomass of the individual sown species. The data for individual samples and the mean over samples for all biomass measures are given. Overall, analyses of the community biomass data have identified species richness as well as functional group composition as important drivers of a positive biodiversity-productivity relationship.

Aboveground plant biomass from the Jena Experiment (Monocultures, year 2003)

This data set contains aboveground plant biomass in 2003 (Sown plant community, Weed plant community, and Dead plant material; all measured in biomass as dry weight) of the monoculture plots of a large grassland biodiversity experiment (the Jena Experiment). In the monoculture plots the biomass of the sown plant community contains only a single species per plot and this species is a different one for each plot. Which species has been sown in which plot is stated in the plot information table for monocultures (see further details below). The monoculture plots of 3.5 x 3.5 m were established for all of the 60 plant species of the Jena Experiment species pool with two replicates per species. These 60 species comprising the species pool of the Jena Experiment belong to four functional groups (grasses, legumes, tall and small herbs). Plots were sown in May 2002 and are since maintained by bi-annual weeding and mowing. Aboveground plant biomass was harvested twice in 2003 just prior to mowing (during peak standing biomass in early June and in late August) on all experimental plots of the monocultures. This was done by clipping the vegetation at 3 cm above ground in 2 rectangles of 0.2 x 0.5 m per plot. The location of these rectangles was assigned prior to each harvest by random selection of coordinates within the core area of the plots (i.e. excluding an outer edge of 0.5 m). The positions of the rectangles within plots were identical for all plots. The harvested biomass was sorted into categories: sown plant species, weed plant species (species not sown at the particular plot), and detached dead plant material (i.e., dead plant material in the data file). All biomass was dried to constant weight (70°C, >= 48 h) and weighed. The data for individual subsamples (i.e. rectangles) and the mean over samples for all biomass measures are given.