Christen D. Børgesen

Changes in time of sowing, flowering and maturity of cereals in Europe under climate change

The phenological development of cereal crops from emergence through flowering to maturity is largely controlled by temperature, but also affected by day length and potential physiological stresses. Responses may vary between species and varieties. Climate change will affect the timing of cereal crop development, but exact changes will also depend on changes in varieties as affected by plant breeding and variety choices. This study aimed to assess changes in timing of major phenological stages of cereal crops in Northern and Central Europe under climate change. Records on dates of sowing, flowering, and maturity of wheat, oats and maize were collected from field experiments conducted during the period 1985–2009. Data for spring wheat and spring oats covered latitudes from 46 to 64°N, winter wheat from 46 to 61°N, and maize from 47 to 58°N. The number of observations (site–year–variety combinations) varied with phenological phase, but exceeded 2190, 227, 2076 and 1506 for winter wheat, spring wheat, spring oats and maize, respectively. The data were used to fit simple crop development models, assuming that the duration of the period until flowering depends on temperature and day length for wheat and oats, and on temperature for maize, and that the duration of the period from flowering to maturity in all species depends on temperature only. Species-specific base temperatures were used. Sowing date of spring cereals was estimated using a threshold temperature for the mean air temperature during 10 days prior to sowing. The mean estimated temperature thresholds for sowing were 6.1, 7.1 and 10.1°C for oats, wheat and maize, respectively. For spring oats and wheat the temperature threshold increased with latitude. The effective temperature sums required for both flowering and maturity increased with increasing mean annual temperature of the location, indicating that varieties are well adapted to given conditions. The responses of wheat and oats were largest for the period from flowering to maturity. Changes in timing of cereal phenology by 2040 were assessed for two climate model projections according to the observed dependencies on temperature and day length. The results showed advancements of sowing date of spring cereals by 1–3 weeks depending on climate model and region within Europe. The changes were largest in Northern Europe. Timing of flowering and maturity were projected to advance by 1–3 weeks. The changes were largest for grain maize and smallest for winter wheat, and they were generally largest in the western and northern part of the domain. There were considerable differences in predicted timing of sowing, flowering and maturity between the two climate model projections applied.

Shifts in comparative advantages for maize, oat and wheat cropping under climate change in Europe

Climate change is anticipated to affect European agriculture, including the risk of emerging or re-emerging feed and food hazards. Indirectly, climate change may influence such hazards (e.g. the occurrence of mycotoxins) due to geographic shifts in the distribution of major cereal cropping systems and the consequences this may have for crop rotations. This paper analyses the impact of climate on cropping shares of maize, oat and wheat on a 50-km square grid across Europe (45–65°N) and provides model-based estimates of the changes in cropping shares in response to changes in temperature and precipitation as projected for the time period around 2040 by two regional climate models (RCM) with a moderate and a strong climate change signal, respectively. The projected cropping shares are based on the output from the two RCMs and on algorithms derived for the relation between meteorological data and observed cropping shares of maize, oat and wheat. The observed cropping shares show a south-to-north gradient, where maize had its maximum at 45–55°N, oat had its maximum at 55–65°N, and wheat was more evenly distributed along the latitudes in Europe. Under the projected climate changes, there was a general increase in maize cropping shares, whereas for oat no areas showed distinct increases. For wheat, the projected changes indicated a tendency towards higher cropping shares in the northern parts and lower cropping shares in the southern parts of the study area. The present modelling approach represents a simplification of factors determining the distribution of cereal crops, and also some uncertainties in the data basis were apparent. A promising way of future model improvement could be through a systematic analysis and inclusion of other variables, such as key soil properties and socio-economic conditions, influencing the comparative advantages of specific crops.

Review of scenario analyses to reduce agricultural nitrogen and phosphorus loading to the aquatic environment.

Nutrient loadings of nitrogen (N) and phosphorus (P) to aquatic environments are of increasing concern globally for managing ecosystems, drinking water supply and food production. There are often multiple sources of these nutrients in the landscape, and the different hydrological flow patterns within stream or river catchments have considerable influence on nutrient transport, transformation and retention processes that all eventually affect loadings to vulnerable aquatic environments. Therefore, in order to address options to reduce nutrient loadings, quantitative assessment of their effects in real catchments need to be undertaken. This involves setting up scenarios of the possible nutrient load reduction measures and quantifying their impacts via modelling. Over the recent two decades there has been a great increase in the use of scenario-based analyses of strategies to combat excessive nutrient loadings. Here we review 130 published papers extracted from Web of Science for 1995 to 2014 that have applied models to analyse scenarios of agricultural impacts on nutrients loadings at catchment scale. The review shows that scenario studies have been performed over a broad range of climatic conditions, with a large focus on measures targeting land cover/use and land management for reducing the source load of N and P in the landscape. Some of the studies considered how to manage the flows of nutrients, or how changes in the landscape may be used to influence both flows and transformation processes. Few studies have considered spatially targeting measures in the landscape, and such studies are more recent. Spatially differentiated options include land cover/use modification and application of different land management options based on catchments characteristics, cropping conditions and climatic conditions. Most of the studies used existing catchment models such as SWAT and INCA, and the choice of the models may also have influenced the setup of the scenarios. The use of stakeholders for designing scenarios and for communication of results does not seem to be a widespread practice, and it would be recommendable for future scenario studies to have a more in-depth involvement of stakeholders for the elaboration and interpretation of scenarios, in particular to enhance their relevance for farm and catchment management and to foster better policies and incentives.

Potential benefits of a spatially targeted regulation based on detailed N-reduction maps to decrease N-load from agriculture in a small groundwater dominated catchment

Denmark must further decrease the N-load to coastal waters from agricultural areas to comply with the Baltic Sea Action Plan and the EU Water Framework Directive. A new spatially targeted regulation is under development that focuses on locating N-mitigation measures in areas with low natural reduction of nitrate (N-reduction). A key tool in this respect is N-reduction maps showing how much N is removed by natural reduction processes, i.e. the ratio between the N-load out of the catchment and the N-leaching from the root zone for each spatial unit within the catchment. For the 85 km2 groundwater dominated Norsminde catchment in Denmark we have analysed the potential benefits of a spatially targeted regulation and how its efficiency is affected by uncertainty in the N-reduction map. Our results suggest that there are potential benefits of implementing a spatially targeted regulation compared to a spatially uniform regulation. The total N-load at the catchment outlet can be decreased up to 8% by relocating the existing agricultural practice according to the N-reduction map and thus without decrease fertilization inputs. A further decrease in N-load can be obtained by identifying target areas with low N-reduction where N-mitigation measures must be applied. Uncertainty on the N-reduction map is found to lower the efficiency of spatially targeted regulation. This uncertainty can be lowered substantially by using the mean of an ensemble of N-reduction maps. The uncertainty decreases with coarser spatial resolution of the N-reduction map, but this will at the same time decrease the benefit from spatially targeted regulation.

Integrated modelling of crop production and nitrate leaching with the Daisy model.

Graphical abstract

Regional greenhouse gas emissions from cultivation of winter wheat and winter rapeseed for biofuels in Denmark

Abstract Biofuels from bioenergy crops may substitute a significant part of fossil fuels in the transport sector where, e.g., the European Union has set a target of using 10% renewable energy by 2020. Savings of greenhouse gas emissions by biofuels vary according to cropping systems and are influenced by such regional factors as soil conditions, climate and input of agrochemicals. Here we analysed at a regional scale the greenhouse gas (GHG) emissions associated with cultivation of winter wheat for bioethanol and winter rapeseed for rapeseed methyl ester (RME) under Danish conditions. Emitted CO2 equivalents (CO2eq) were quantified from the footprints of CO2, CH4 and N2O associated with cultivation and the emissions were allocated between biofuel energy and co-products. Greenhouse gas emission at the national level (Denmark) was estimated to 22.1 g CO2eq MJ−1 ethanol for winter wheat and 26.0 g CO2eq MJ−1 RME for winter rapeseed. Results at the regional level (level 2 according to the Nomenclature of Territorial Units for Statistics [NUTS]) ranged from 20.0 to 23.9 g CO2eq MJ−1 ethanol and from 23.5 to 27.6 g CO2eq MJ−1 RME. Thus, at the regional level emission results varied by up to 20%. Differences in area-based emissions were only 4% reflecting the importance of regional variation in yields for the emission result. Fertilizer nitrogen production and direct emissions of soil N2O were major contributors to the final emission result and sensitivity analyses showed that the emission result depended to a large extent on the uncertainty ranges assumed for soil N2O emissions. Improvement of greenhouse gas balances could be pursued, e.g., by growing dedicated varieties for energy purposes. However, in a wider perspective, land-use change of native ecosystems to bioenergy cropping systems could compromise the CO2 savings of bioenergy production and challenge the targets set for biofuel production.

Spatially differentiated strategies for reducing nitrate loads from agriculture in two Danish catchments

Nutrient loss from agriculture is the largest source of diffuse water pollution in Denmark. To reduce nutrient loads a number of solutions have been implemented, but this has been insufficient to achieve the environmental objectives without unacceptable repercussions for agricultural production. This has substantiated the need to develop a new approach to achieve nitrogen (N) load reduction to the aquatic environments with lower costs to farmers. The new approach imply targeting N leaching mitigation to those parts of the landscape which contribute most to the N-loadings. This would involve either reducing the source loading or enhancing the natural reduction (denitrification) of N after it is leached from the root zone of agricultural crops. In this study, a new method of spatially differentiated analysis for two Danish catchments (Odense and Norsminde) was conducted that reach across the individual farms to achieve selected N-load reduction targets. It includes application of cover crops within current crop rotations, set-a-side application on high N-load areas, and changes in agricultural management based on maps of N-reduction available for two different spatial scales, considering soil type and farm boundaries as spatial constraints. In summary, the results revealed that considering spatial constraints for changes in agricultural management will affect the effectiveness of N-load reduction, and the highest N-load reduction was achieved where less constraints were considered. The results also showed that the range of variation in land use, soil types, and N-reduction potential influence the reduction of N-loadings that can originate from critical source areas. The greater the spatial variation the greater the potential for N load reduction through targeting of measures. Therefore, the effectiveness of spatially differentiated measures in term of set-a-side area in Odense catchment were relatively greater compared to Norsminde catchment. The results also showed that using a fine spatial N-reduction map provides greater potential for N load reductions compared to using sub-catchment scale N-reduction maps.

Potential benefits of farm scale measures versus landscape measures for reducing nitrate loads in a Danish catchment

To comply with the EU Water Framework Directive, Denmark must further reduce the nitrate (N)-load to marine ecosystems from agricultural areas. Under the anticipated future spatially targeted regulation, the required N-load reductions will differ between catchments, and these are expected to be mitigated by a combination of land and water management measures. Here, we explored how the expected N-load reduction target of 38% for a Danish catchment (River Odense) could be achieved through a combination of farm and landscape measures. These include: (a) N-leaching reduction through changing the crop rotation and applying cover crops, (b) enhancing N-reduction through (re)establishment of wetlands, and (c) reducing N-leaching through spatially targeting of set-aside to high N-load areas. Changes in crop rotations were effective in reducing N-leaching by growing crops with a longer growing season and by allowing a higher use of cover crops. A combination of wetlands and changes in crop rotations were needed for reaching the N-load reduction target without use of set-aside. However, not all combinations of wetlands and crop rotation changes achieved the required N-load reduction, resulting in a need for targeted set-aside, implying a need for balancing measures at farm and landscape scale to maximize N load reduction while minimizing loss of productive land. The effectiveness of farm scale measures is affected by farm and soil types as well as by N-reduction in groundwater, while the possibilities for using wetlands for decreasing the N-load depends on landscape features, allowing the establishment of wetlands connected to streams and rivers.

Delay in catchment nitrogen load to streams following restrictions on fertilizer application

A MIKE SHE hydrological-solute transport model including nitrate reduction is employed to evaluate the delayed response in nitrogen loads in catchment streams following the implementation of nitrogen mitigation measures since the 1980s. The nitrate transport lag times between the root zone and the streams for the period 1950-2011 were simulated for two catchments in Denmark and compared with observational data. Results include nitrogen concentration and mass discharge to streams. By automated baseflow separation, stream discharge was separated into baseflow and drain flow components, and the nitrogen concentration and mass discharge in baseflow and drain flow were determined. This provided insight on the development of stream nitrogen loads, with a short average lag time in drain flow and a long average lag time in baseflow. The long term effect of nitrogen mitigation measures was determined, with results showing that there is a 15 years long delay in the appearance of peak nitrogen loads in streams. This means that real time stream monitoring data cannot be used alone to assess the effect of nitrogen mitigation measures.

Modelling agro-environmental variables under data availability limitations and scenario managements in an alluvial region of the North China Plain

Abstract Single or multiple weather station data were combined with soil textural data ranging from low to high detail, i.e., point data from a field station, the FAO Digital Soil Map of the World and a comprehensive data from national soil survey, as input to the Daisy model to simulate and upscale crop yields, drainage and nitrogen leaching for an agroecosystem in the North China Plain. Increasing the detail of the weather data increased the spatial variation of all simulated variables and decreased their regional median. Regional crop yields were simulated well with high-detail input data, though at a weak response to data detail. Simulated regional drainage and nitrogen leaching, and their spatial variability, however, responded well and increased two-to threefold, but their regional medians were similar for medium- and high-detail soil data. This work demonstrates the importance of explicit consideration of weather and soil variability for agro-environmental simulation studies at regional scale.