Risk/opportunity:(from the Climate Change Risk Assessment for Scotland 2012):
AG1b/d Changes in wheat and spring barley yield (due to warmer springs AG1c Changes in potato yield AG1e Changes in winter barley yield (due to wetter winters)

Narratives: Suitability and productivity (agriculture)

SCCAP theme: Natural environment

SCCAP objectives:
N1: Understand the effects of climate change and their impacts on natural environment
N3: Sustain and enhance the benefits, goods and services that the natural environment provides

Latest figures

Due to a high solar income but relatively little water stress during the summer of 2014, yields (for winter wheat, spring barley and oilseed rape) were correspondingly high (though not significantly higher than had been recorded previously in the last decade.

Trend
At a glance
  • The future trends of yield and inputs are difficult to predict because of some major uncertainties that are not directly weather-related.
  • The highest input crops such as potato, winter wheat and winter oilseed rape are susceptible to unusual or extreme weather.
  • Farmers need to consider a variety of strategies to reduce the sensitivity of their crops to ‘out-of-the-ordinary weather’: reduce reliance on more sensitive crops or varieties, alter cropping systems, improve soil condition.

The cereal-based croplands of the east of Scotland are high yielding by UK standards because of a well-established phenomenon in which crops bulk large quantities of carbon in the long, cool summer of northern latitudes. Between March and September, and compared to more southerly parts of the UK, the lower temperature in the north stretches out the development of the crops, with the result that they are in the field for longer and thereby able to capture a large fraction of the incoming solar energy[1].

Variation in weather from year to year has two types of effect. First, it alters the phases of crop development leading to less or more solar energy being captured and converted to plant matter. Second, an unusual event can interfere with crucial stages of the crops’ cycle: for example, wet weather at flowering can limit pollination in some crops, while prolonged wet in late summer can cause cereal grain to germinate and rot while still on the plant and cause disease and harvesting problems in potato.

Yield itself is of limited use as an indicator, since it depends on the interactions between the weather, soil condition and agronomic inputs.  Change in the latter two can affect yield as least as much as between-year variation in the weather. Therefore an indicator needs to consider both yield itself and agronomic inputs. Among the latter only fertiliser and pesticide have been recorded annually with enough spatial coverage to be representative of Scotland[2]. The indicator set therefore consists of three primary indicators and two ‘efficiencies’ calculated from the primary:

  • yield of crops by type, stated to a specified water content (t ha-1)
  • fertiliser used on crops, mainly nitrogen, phosphate and potash (kg ha-1)
  • pesticide used on crops, either as fungicide, herbicide, insecticide, etc. presented as (for example) mass of active ingredient or ‘spray area index’ (=area sprayed with pesticide/area sown with crop)
  • fertiliser ‘efficiency’ - yield per unit fertiliser
  • pesticide ‘efficiency’ - yield per unit pesticide.

Related indicators:

NA28 Wetness risk for agriculture (arable suitability and grassland suitability)

NA29 Drought risk to agricultural land


[1] The incoming solar, measured as a long term average, is about 10% less than in the south, but days are longer in the north and the incoming solar flux per unit area per unit time is lower, and as a consequence, evaporation rate and leaf temperature are generally lower in summer and more favourable to growth. The allocation of dry matter between vegetative parts and yield (grain or tuber) tends to be conservative such that yield is often a linear function of intercepted solar energy.

[2] Several other facets of agronomic input are not available simply because they have not been systematically measured, among which are descriptors of tillage, machinery and traffic over fields.

The steep upward trajectory of yield after the 1960s (see next section) levelled in the 1990s for most crops. The exception perhaps is winter oilseed rape which has shown a steady though small upward trajectory since 2000. Similarly, fertiliser input stabilised in arable crops in the mid-1990s and has varied up and down over the years since. In contrast, nitrogen applied to grass followed a consistent decline during and after the 1990s. Pesticides in general have continued to increase in terms of the number of different applications applied to most types of crop.

The stability of yields over the past 15 years provides a reference or background for the ‘bad’ year of 2012 and the ‘good’ year of 2014. Yields are shown for three crops in Fig. 1a. Winter wheat is the highest yielding cereal, producing more than spring barley because it is in the ground for longer and has enough leaf to capture the increasing solar energy in spring and early summer. Winter oilseed rape has a lower mass-yield than the cereals, partly due to the higher ’calorific content’ of the oil in the seed and partly to the lower harvestable fraction, i.e. harvested seed is proportionately less of the whole plant than in the cereals. Potato (not shown) has a higher dry matter yield because a larger fraction of the total plant matter is allocated to the tubers than to grain in the cereals.

Figure 1 a) Yield of winter wheat (WW), spring barley (SB) and winter oilseed rape (WOSR) between 2002 and 2014 from Scottish Government statistics. b) The same yield data expressed as an ‘efficiency’ defined as yield/ nitrogen input (note the reversal of WW and SB).

The very wet year, 2012, caused yield of winter wheat, oilseed rape and potato to drop by 30% or more. However, the more resilient crop, spring barley, was less sensitive to that particular year, though the physiological reasons why crops differed are not fully understood. In contrast, the summer of 2014 was characterised by a high solar income but relatively little water stress and yields were correspondingly high (Fig. 1a), though only just or no higher than had been recorded previously in the last decade.Figure 1 a) Yield of winter wheat (WW), spring barley (SB) and winter oilseed rape (WOSR) between 2002 and 2014 from Scottish Government statistics. b) The same yield data expressed as an ‘efficiency’ defined as yield/ nitrogen input (note the reversal of WW and SB).

When expressed as an ‘efficiency’ (yield / nitrogen input), spring barley comes out on top, producing more grain than winter wheat for each kg of nitrogen applied (Fig. 1b). This is a consequence of the ‘diminishing returns’ of increasing intensification (see argument below). Moreover, all three crops appear to show a higher efficiency in the centre of the range of years in Fig. 1b. The cause of the variation lies in the nitrogen rather than the yield, and is due to slightly less fertiliser being used. There appears to have been a re-adjustment towards the end of the decade 2000-2010, since the efficiency started to move back down to the earlier values, but any short term trend was obscured by the wet year of 2012. Since farmers did not anticipate the wet summer and autumn of 2012, they applied typical fertiliser, such that a large drop in efficiency occurred in the cereals that year.

The wet year of 2012 and good growing weather of 2014[1] both indicate the degree of sensitivity of crops to unusual weather. But any impacts of trends in weather and climate will be difficult to distinguish from other factors, such as:

  • short term trends in increased use or reduced use of fertiliser (perhaps in compensation for a previous period of over-use or under-use) as perhaps in Fig. 1b;
  • trends and variation in the quantity, timing and type of pesticide;
  • yield nearing a ‘ceiling’ for the region, such that increases due to ‘good’ weather will be constrained whereas decreases due to bad weather will not;

soil degradation due to continued intensification (Squire et al. 2015).


[1] not yet fully analysed, nitrogen data not available

The recent period of intensification after the 1960s was accompanied by large changes in inputs to and outputs from agriculture. The available information is not all complete but main trends are discernible in some indicators, notably total fertiliser used in Scotland (measured in kt) - nitrogen showing a major increase after the early 1960s and phosphate a systematic fall in usage to the present (Fig. 2). However, the ‘rate’ of application (kg ha-1) to fields growing different crops was apparently not recorded systematically in Scotland before 1982 and the first pesticide use survey was in 1974.

Figure 2 Total fertiliser applied in Scotland to all crops and grass: closed circles phosphate, open triangles nitrogen. Source: Defra, see References.

Yields of the main crops began an upward trajectory in the 1960s and 1970s (depending on the crop) due to intensification during which yield per unit area and inputs both increased several-fold (Fig. 3).

The main types of intensification were:

  • spring-sown crops received more fertiliser (particularly nitrogen) and pesticide;
  • winter (autumn-sown) crops notably winter wheat received much more fertiliser and pesticide, with great effect, causing yields of wheat and spring barley to diverge from similar values in the early 1970s, until wheat produced almost twice the yield of spring barley  by the early 1990s;
  • new winter crops came into the system, particular winter oilseed rape which received more nitrogen fertiliser than even winter wheat;
  • high intensity winter crops replaced spring crops in some regions, thereby increasing overall inputs;
  • most other crops, notably potato, received more fertiliser and pesticide;
  • mechanisation and the size of farm machinery increased.

The greatest changes occurred over about fifteen years between the late 1970s and early 1990s. Subsequently, some inputs per unit field area stabilised or even decreased for a time, for example nitrogen fertiliser, while others continued to increase, particularly the number and different types of pesticide application.

Figure 3 Trends in grain yield (t ha-1) and pesticide area index[1] for wheat, mainly winter varieties (closed square, solid line) and spring barley (open square, dashed line). Adapted from Squire et al. (2015)

However, unravelling the long-term trends in fertiliser applied to farmland and the consequent effects on yield presents challenges. In particular, the ‘rate’ of application of the different fertilisers in kg ha-1 to the different crops, which is essential for analysing and explaining trends, is not available throughout the period of intensification[2].

Since, however, the various crops differ considerably in the quantities of fertiliser they receive, grouping them all together as ‘tillage crops’ is misleading. Data for individual crops after 1982 are available on request and the more recent of these data are accessible online, but the omission of information on fertiliser inputs to individual crops before 1982 in Scotland is an obstacle to the analysis of long-term trends.

The first pesticide use survey, in 1974, was near the start of the main phase of intensification. There followed occasional surveys up to the mid-1990s then surveys on arable crops every two years. Throughout this period, the yield of the crops per unit pesticide applied declined greatly, largely due to the continued rise in the number of pesticide formulations given to crops, despite the rise in yield slowing down and finally stopping for most crops (compare changes in yield and pesticide in Fig. 3)[3]. Further work is needed to explore yield-efficiencies in relation to various attributes of the pesticides.


[1] Pesticide Area Index is calculated from Pesticide Use data by summing the area treated with the different types of pesticide (mainly fungicide, herbicide, insecticide) and dividing the total by the area sown with the crop.

[2] Fertiliser Practice (Defra) gives annual total quantities applied in kt for Scotland (e.g. Fig. 2), while application rates (kg ha-1) for tillage crops and grass, but not individual crops, are given separately only from 1982 onwards. To include the earlier period, the data for England and Wales before 1982 (e.g. Church & Lewis, 1977), can be used as a guide, i.e. the values for E&W multiplied by the mean ratio for Scotland to E&W after 1982.

[3] The values of efficiency, expressed as yield/pesticide, depend on how pesticides are themselves quantified. The index used here is similar to the previously used ‘spray area index’ and depends on the number of formulations applied to crops without taking account of the ecotoxicity of the chemicals.

The future trends of yield and inputs are difficult to predict because of some major uncertainties that are not directly weather-related. These include:

  • the levelling of yield for which there are several possible reasons (Knight et al. 2012)
  • the contribution of degradation of soil  due to high-intensity farming (Squire et al. 2015)
  • lack of information on effects of other change e.g. tillage practices.

If nothing else changes, then warmer springs should encourage timely sowing of the spring crops, more rapid emergence and more rapid early development. The crops will move through the vegetative stage quicker, but the result will depend on the timing of the annual solar and temperature cycles and on the innate determinacy of the crop[1].

The effects of wetter winters on winter cereals and oilseed rape will depend on the severity of the conditions, but the effects on crops of different degrees of wet and cold in the field are not well researched and so difficult to predict.

Another topic of interest is whether heavy machinery can be driven on a field. Typically, winter crops are treated with fertiliser and pesticide in the spring but timing can vary substantially. Wetter winters will generally restrict the time that field can be accessed. But as indicated earlier, the consequence may depend on the degree to which current and previous practice has already degraded soil condition.

In general, terms like warmer springs and wetter winters are too coarse to enable any precise prediction of future trends. Yet even if a more detailed scenario was presented and considered, the lack of knowledge regarding the effects of factors such as waterlogging on crops would mean that responses would be hard to define. What is clear is that today’s highest input crops such as potato, winter wheat and winter oilseed rape, are sensitive to out-of-the-ordinary weather. It seems little attention is given by seed companies to new crop varieties that might overcome such conditions and deliver a more stable yield over time, especially since Scotland is a not a prime target region for most companies. The most productive approach towards achieving a less sensitive yield may be -

  • to search among current crop varieties for those that are least sensitive to unusual runs of weather;
  • to devise cropping systems, e.g. based on varietal mixtures and mixed-species crops, that are more stable and resilient;
  • to phase out practices, such as current methods of tillage in potato fields, that are highly damaging to soil and instead introduce practices that result in improved soil condition;
  • rely less on sensitive crops by devising a wider range of economic products from the less sensitive crops such as spring barley.

[1] A determinate, reproductive crop such as a cereal may intercept less solar radiation in its vegetative phase, accrue less dry matter and so have less to retranslocate from stem to grain later in the year. The same may not be true of the more indeterminate, vegetative potato, where faster early development should lead to greater accumulation of mass throughout vegetative growth which continues for much of the season.

Regional variation exists in the proportions of the main crops grown. There are various reasons including:

  • a crop is linked with other uses that have a regional emphasis – for example spring barley grown for animal feed on the same farm;
  • wheat preferentially grown on soils of higher clay, potato on soils of lower clay.

So for example there is greater ratio of wheat to spring barley in the Lothians, but a smaller ratio in parts of Aberdeenshire. It should be possible to get data for yield, pesticide and fertiliser broken down by region or parish-groupings, but from the three different sources. This was not attempted in the present study.

The main upward trends in yield over the period 1960-1990 were driven largely by greater inputs of nitrogen and pesticide, new crop varieties that responded to those inputs and general improvements in mechanisation and agronomic practice (many of the latter not systematically recorded). The factors responsible for the slowing down of the rise and then the levelling in yield after the mid-1990s are difficult to identify for the following reasons.

  • many variables were changing at the same time – agronomy, crop varieties, soil condition, pests and pest control, weather;
  • actual yields may have been moving towards the potential yield or ceiling defined by climate and soil, such that any single variable promoting a rise in yield will have limited impact;
  • if upward movements in yield are limited by nearness to a ceiling, any positive effects on yield of climate or weather may be hardly realised whereas negative effects will be.

Research effort now needs to be directed towards estimating both the climatic ceiling, which  depends on factors such as solar income, temperature and evaporation, and the actual (lower) ceiling imposed by combinations of climate, soil and biotic factors such as pests. It might be argued that neither ceiling has been reached because much higher yields than the average are produced in some years by some farmers and new crop varieties appear to yield more in trials than the current ones. But the main indicators used here are averages, and it may be that while combinations of certain fields, farmers, crop varieties and years will return a high yield, the arable surface as a whole is nearing the limit of its production. Further investigations are needed to establish the position.

The main limitations to the analysis of existing census data are

  • yield, fertiliser and pesticide are all collected by different methods and different organisations and over different averaging scales;
  • comparing averages of the three could be misleading - progress might lie in finding the best combination of inputs and agronomy which may be very different from the averages;
  • therefore information per farm or per field would allow much more insight as to whether certain farmers are able to adapt to or overcome limitations due to climate (but government sources tend not to release information at these scales);and
  • in conclusion, much greater attention and effort is needed by the various agencies to standardise their data collection such that the yield, fertiliser and pesticide data are all derived from the same fields and farms.

Church BM, Lewis DA. 1977 Fertiliser use on farm crops, England and Wales: information from the Survey of Fertiliser Practice, 1942-1976. Outlook Agr. 9, 186-193.

Economic Report on Scottish Agriculture. 2013 (and previous annual reports in this series). Edinburgh: RESAS, Scottish Government. http://www.scotland.gov.uk/Publications/2014/06/3709/14#ii

Fertiliser Practice. 2013 The British survey of fertiliser practice: fertiliser use on farm crops for crop year 2012 (and previous reports in this series). London: Defra. https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/192605/fertiliseruse-report2012-25apr13.pdf

Scottish Government 2014. Final estimate of the cereal and oilseed rape harvest 2014. http://www.scotland.gov.uk/Publications/2014/12/2462 

Knight S, et al. 2012. Desk study to evaluate contributory causes of the current ‘yield plateau’ in wheat and oilseed rape. HGCA/Defra.

Squire GR, Hawes C, Valentine TA, Young MW (2015) Degradation rate of soil function varies with trajectory of agricultural intensification. Agriculture, Ecosystems and Environment (in press).

Watson J, Hughes J, Thomas L, Wardlaw J. 2013 Pesticide usage in Scotland. Arable Crops 2012. SASAS Edinburgh

Primary author: G R Squire, James Hutton Institute