Filtern
Erscheinungsjahr
Dokumenttyp
Gehört zur Bibliographie
- ja (51) (entfernen)
Schlagworte
- Principal component analysis (4)
- evapotranspiration (4)
- Nitrogen (3)
- Water quality (3)
- wetland (3)
- CORDEX data (2)
- Carbon (2)
- Catchment classification (2)
- Climate change (2)
- Drought indices (2)
Provisioning a sufficient stable source of food requires sound knowledge about current and upcoming threats to agricultural production. To that end machine learning approaches were used to identify the prevailing climatic and soil hydrological drivers of spatial and temporal yield variability of four crops, comprising 40 years yield data each from 351 counties in Germany. Effects of progress in agricultural management and breeding were subtracted from the data prior the machine learning modelling by fitting smooth non-linear trends to the 95th percentiles of observed yield data. An extensive feature selection approach was followed then to identify the most relevant predictors out of a large set of candidate predictors, comprising various soil and meteorological data. Particular emphasis was placed on studying the uniqueness of identified key predictors. Random Forest and Support Vector Machine models yielded similar although not identical results, capturing between 50% and 70% of the spatial and temporal variance of silage maize, winter barley, winter rapeseed and winter wheat yield. Equally good performance could be achieved with different sets of predictors. Thus identification of the most reliable models could not be based on the outcome of the model study only but required expert's judgement. Relationships between drivers and response often exhibited optimum curves, especially for summer air temperature and precipitation. In contrast, soil moisture clearly proved less relevant compared to meteorological drivers. In view of the expected climate change both excess precipitation and the excess heat effect deserve more attention in breeding as well as in crop modelling.
Local biodiversity patterns are expected to strongly reflect variation in topography, land use, dispersal boundaries, nutrient supplies, contaminant spread, management practices, and other anthropogenic influences. Contrary to this expectation, studies focusing on specific taxa revealed a biodiversity homogenization effect in areas subjected to long-term intensive industrial agriculture. We investigated whether land use affects biodiversity levels and community composition (alpha- and beta-diversity) in 67 kettle holes (KH) representing small aquatic islands embedded in the patchwork matrix of a largely agricultural landscape comprising grassland, forest, and arable fields. These KH, similar to millions of standing water bodies of glacial origin, spread across northern Europe, Asia, and North America, are physico-chemically diverse and differ in the degree of coupling with their surroundings. We assessed aquatic and sediment biodiversity patterns of eukaryotes, Bacteria, and Archaea in relation to environmental features of the KH, using deep-amplicon-sequencing of environmental DNA (eDNA). First, we asked whether deep sequencing of eDNA provides a representative picture of KH aquatic biodiversity across the Bacteria, Archaea, and eukaryotes. Second, we investigated if and to what extent KH biodiversity is influenced by the surrounding land use. We hypothesized that richness and community composition will greatly differ in KH from agricultural land use compared with KH in grasslands and forests. Our data show that deep eDNA amplicon sequencing is useful for in-depth assessments of cross-domain biodiversity comprising both micro- and macro-organisms, but has limitations with respect to single-taxa conservation studies. Using this broad method, we show that sediment eDNA, integrating several years to decades, depicts the history of agricultural land-use intensification. Aquatic biodiversity was best explained by seasonality, whereas land-use type explained little of the variation. We concluded that, counter to our hypothesis, land use intensification coupled with landscape wide nutrient enrichment (including atmospheric deposition), groundwater connectivity between KH and organismal (active and passive) dispersal in the tight network of ponds, resulted in a biodiversity homogenization in the KH water, leveling off today's detectable differences in KH biodiversity between land-use types. These findings have profound implications for measures and management strategies to combat current biodiversity loss in agricultural landscapes worldwide.
Land-use type temporarily affects active pond community structure but not gene expression patterns
(2022)
Changes in land use and agricultural intensification threaten biodiversity and ecosystem functioning of small water bodies. We studied 67 kettle holes (KH) in an agricultural landscape in northeastern Germany using landscape-scale metatranscriptomics to understand the responses of active bacterial, archaeal and eukaryotic communities to land-use type. These KH are proxies of the millions of small standing water bodies of glacial origin spread across the northern hemisphere. Like other landscapes in Europe, the study area has been used for intensive agriculture since the 1950s. In contrast to a parallel environmental DNA study that suggests the homogenization of biodiversity across KH, conceivably resulting from long-lasting intensive agriculture, land-use type affected the structure of the active KH communities during spring crop fertilization, but not a month later. This effect was more pronounced for eukaryotes than for bacteria. In contrast, gene expression patterns did not differ between months or across land-use types, suggesting a high degree of functional redundancy across the KH communities. Variability in gene expression was best explained by active bacterial and eukaryotic community structures, suggesting that these changes in functioning are primarily driven by interactions between organisms. Our results indicate that influences of the surrounding landscape result in temporary changes in the activity of different community members. Thus, even in KH where biodiversity has been homogenized, communities continue to respond to land management. This potential needs to be considered when developing sustainable management options for restoration purposes and for successful mitigation of further biodiversity loss in agricultural landscapes.
The curse of the past
(2021)
One challenge for modern agricultural management schemes is the reduction of harmful effects on the envi-ronment, e.g. in terms of the emission of nutrients. Sampling the effluent of tile drains is a very efficient way to sample seepage water from larger areas directly underneath the main rooting zone. Time series of solute con-centration in tile drains can be linked to agricultural management data and thus indicate the efficacy of individual management measures. To that end, the weekly runoff and solute concentration were determined in long-term measurement campaigns at 25 outlets of artificial tile drains at 19 various arable fields in the German federal state of Mecklenburg-Vorpommern. The study sites were distributed within a 23,000 km(2) region and were deemed representative of intense arable land use. In addition, comprehensive meteorological and man-agement data were provided. To disentangle the different effects, monitoring data were subjected to a principal component analysis. Loadings on the prevailing principal components and spatial and temporal patterns of the component scores were considered indicative of different processes. Principal component scores were then related to meteorological and management data via random forest modelling. Hydrological conditions and weather were identified as primary driving forces for the nutrient discharge behaviour of the drain plots, as well as the nitrogen balance. In contrast, direct effects of recent agricultural management could hardly be identified. Instead, we found clear evidence of the long-term and indirect effects of agriculture on nearly all solutes. We conclude that tile drain effluent quality primarily reflected the soil-internal mobilisation or de-mobilisation of nutrients and related solutes rather than allowing inferences to be drawn about recent individual agricultural management measures. On the other hand, principal component analysis revealed a variety of indirect and long-term effects of fertilisation on solutes other than nitrogen or phosphorus that are still widely overlooked in nutrient turnover studies.
This study investigates possible impacts of four global warming levels (GWLs: GWL1.5, GWL2.0, GWL2.5, and GWL3.0) on drought characteristics over Niger River basin (NRB) and Volta River basin (VRB). Two drought indices-Standardized Precipitation Index (SPI) and Standardized Precipitation-Evapotranspiration Index (SPEI)-were employed in characterizing droughts in 20 multi-model simulation outputs from the Coordinated Regional Climate Downscaling Experiment (CORDEX). The performance of the simulation in reproducing basic hydro-climatological features and severe drought characteristics (i.e., magnitude and frequency) in the basins were evaluated. The projected changes in the future drought frequency were quantified and compared under the four GWLs for two climate forcing scenarios (RCP8.5 and RCP4.5). The regional climate model (RCM) ensemble gives a realistic simulation of historical hydro-climatological variables needed to calculate the drought indices. With SPEI, the simulation ensemble projects an increase in the magnitude and frequency of severe droughts over both basins (NRB and VRB) at all GWLs, but the increase, which grows with the GWLs, is higher over NRB than over VRB. More than 75% of the simulations agree on the projected increase at GWL1.5 and all simulations agree on the increase at higher GWLs. With SPI, the projected changes in severe drought is weaker and the magnitude remains the same at all GWLs, suggesting that SPI projection may underestimate impacts of the GWLs on the intensity and severity of future drought. The results of this study have application in mitigating impact of global warming on future drought risk over the regional water systems.
Large-scale crop yield failures are increasingly associated with food price spikes and food insecurity and are a large source of income risk for farmers. While the evidence linking extreme weather to yield failures is clear, consensus on the broader set of weather drivers and conditions responsible for recent yield failures is lacking. We investigate this for the case of four major crops in Germany over the past 20 years using a combination of machine learning and process-based modelling. Our results confirm that years associated with widespread yield failures across crops were generally associated with severe drought, such as in 2018 and to a lesser extent 2003. However, for years with more localized yield failures and large differences in spatial patterns of yield failures between crops, no single driver or combination of drivers was identified. Relatively large residuals of unexplained variation likely indicate the importance of non-weather related factors, such as management (pest, weed and nutrient management and possible interactions with weather) explaining yield failures. Models to inform adaptation planning at farm, market or policy levels are here suggested to require consideration of cumulative resource capture and use, as well as effects of extreme events, the latter largely missing in process-based models. However, increasingly novel combinations of weather events under climate change may limit the extent to which data driven methods can replace process-based models in risk assessments.
Natural ponds are perceived as spatially and temporally highly variable ecosystems. This perception is in contrast to the often-applied sampling design with high spatial but low temporal replication. Based on a data set covering a period of six years and 20 permanently to periodically inundated ponds, we investigated whether this widely applied sampling design is sufficient to identify differences between single ponds or single years with regard to water quality and macrophyte community composition as measures of ecosystem integrity.
In our study, the factor "pond", which describes differences between individual ponds, explained 56 % and 63 %, respectively, of the variance in water quality and macrophyte composition. In contrast, the factor "year" that refers to changes between individual years, contributed less to understand the observed variability in water quality and macrophyte composition (10 % and 7 % respectively, of the variance explained). The low explanation of variance for "year" and the low year-to-year correlation for the single water quality parameter or macrophyte coverage values, respectively, indicated high but non-consistent temporal variability affecting individual pond patterns.
In general, the results largely supported the ability of the widely applied sampling strategy with about one sampling date per year to capture differences in water quality and macrophyte community composition between ponds. Hence, future research can be rest upon sampling designs that give more weight to the number of ponds than the number of years in dependence on the research question and the available resources. Nonetheless, pond research would miss a substantial amount of information (7 to 10 % of the variance explained), when the sampling would generally be restricted to one year. Moreover, we expect that the importance of multiple-year sampling will likely increase in periods and regions of higher hydrological variability compared to the average hydrological conditions encountered in the studied period.
This study investigates possible impacts of four global warming levels (GWLs: GWL1.5, GWL2.0, GWL2.5, and GWL3.0) on drought characteristics over Niger River basin (NRB) and Volta River basin (VRB). Two drought indices-Standardized Precipitation Index (SPI) and Standardized Precipitation-Evapotranspiration Index (SPEI)-were employed in characterizing droughts in 20 multi-model simulation outputs from the Coordinated Regional Climate Downscaling Experiment (CORDEX). The performance of the simulation in reproducing basic hydro-climatological features and severe drought characteristics (i.e., magnitude and frequency) in the basins were evaluated. The projected changes in the future drought frequency were quantified and compared under the four GWLs for two climate forcing scenarios (RCP8.5 and RCP4.5). The regional climate model (RCM) ensemble gives a realistic simulation of historical hydro-climatological variables needed to calculate the drought indices. With SPEI, the simulation ensemble projects an increase in the magnitude and frequency of severe droughts over both basins (NRB and VRB) at all GWLs, but the increase, which grows with the GWLs, is higher over NRB than over VRB. More than 75% of the simulations agree on the projected increase at GWL1.5 and all simulations agree on the increase at higher GWLs. With SPI, the projected changes in severe drought is weaker and the magnitude remains the same at all GWLs, suggesting that SPI projection may underestimate impacts of the GWLs on the intensity and severity of future drought. The results of this study have application in mitigating impact of global warming on future drought risk over the regional water systems.
Groundwater levels are monitored by environmental agencies to support the sustainable use of groundwater resources. For this purpose continuous and spatially comprehensive monitoring in high spatial and temporal resolution is desired. This leads to large datasets that have to be checked for quality and analysed to distinguish local anthropogenic influences from natural variability of the groundwater level dynamics at each well. Both technical problems with the measurements as well as local anthropogenic influences can lead to local anomalies in the hydrographs. We suggest a fast and efficient screening method for the identification of well-specific peculiarities in hydrographs of groundwater head monitoring networks. The only information required is a set of time series of groundwater heads all measured at the same instants of time. For each well of the monitoring network a reference hydrograph is calculated, describing expected "normal" behaviour at the respective well as is typical for the monitored region. The reference hydrograph is calculated by multiple linear regression of the observed hydrograph with the "stable" principal components (PCs) of a principal component analysis of all groundwater head series of the network as predictor variables. The stable PCs are those PCs which were found in a random subsampling procedure to be rather insensitive to the specific selection of the analysed observation wells, i.e. complete series, and to the specific selection of measurement dates. Hence they can be considered to be representative for the monitored region in the respective period. The residuals of the reference hydrograph describe local deviations from the normal behaviour. Peculiarities in the residuals allow the data to be checked for measurement errors and the wells with a possible anthropogenic influence to be identified. The approach was tested with 141 groundwater head time series from the state authority groundwater monitoring network in northeastern Germany covering the period from 1993 to 2013 at an approximately weekly frequency of measurement.