Soil Erosion, Climate Change and Global Food Security: Challenges and Strategies. Part 4.

This is the fourth part of a much longer article published in the journal Science Progress, and which may be found here: http://stl.publisher.ingentaconnect.com/content/stl/sciprg/2014/00000097/00000002/art00001


9. Establishing a relationship between land degradation, soil productivity and crop yields.

The productivity of some lands has fallen by 50% as a result of soil erosion and desertification, and the according reduction in crop yields in Africa lies in the range 2─40%, with a mean loss of 8.2% for the continent overall11. The loss of productivity in South Asia has been reckoned at 36 million tons of cereal equivalent with a value of $5.4 billion as a result of water erosion and $1.8 billion from wind erosion11. It is a vexed matter to make a definite connection between the extent and processes of soil erosion and declining crop yields, since the latter may result from various influences. In some cases, the crop yields do not fall markedly, and may even increase for a time, despite the soil being eroded, e.g. if a compensatory increase is made in fertilizer inputs. Crop yields may be impaired43 by an excessive removal of nutrients from the soil, which are not replenished; the impact of pests and diseases; weed infestations; and the greater frequency of drought as a consequence of climate change. Other factors – which may be associated with soil erosion – can also be culpable for a reduction in crop yields, e.g. a restriction in the possible rooting depth, e.g. when the soil depth becomes limited, and the roots touch the bedrock or a clay layer; a reduction in the water capacity of the soil; a decline in SOM (SOC) content; an increasing salinity or sodicity of soil; other changes in the chemical composition of the soil: e.g. the presence of aluminium or heavy metal cations, or a reduction in soil pH (acidification) in general. All of the above, in one way or another, are connected to some type of soil degradation, the most common being soil erosion by water43.

It can be said that practically any adverse environmental change is likely to lead to soil erosion and a decrease in biomass yield, such is the inextricable complexity of the underlying components of these phenomena. To invoke a spectrum of impact, we may at one extreme consider the conversion of dryland savannah to continuous cropping (the practice of growing the same crop in the same space year after year) of soya beans (soybeans). As a result of this change, the combined influence of loss of vegetation cover and soil disturbance will aggravate and accelerate soil erosion. Although the crop (a legume) will contribute some nitrogen to the soil, and some organic matter, a tipping point will ensue eventually when production is impaired, as a result of a thinning of the topsoil, colloid loss, and a reduction in the water-retaining capacity of the soil. However, the input of resources (e.g. fertilizers, irrigation) and technological means can allow production to be continued unabated. At the other end of this spectrum are the “badlands” – a result of the mistreatment of semiarid ground, where serious soil erosion has occurred, with gullies, rills, pipes and other related aspects – which are completely lacking in vegetation. As far as apportioning blame for the loss of vegetation to soil erosion is concerned, both extremes are really “chicken and egg” situations: erosion must result in a reduced soil quality, which impairs plant production and reproduction - allowing that this might be masked by technology and other inputs – but at the same time, a loss of vegetative cover provokes soil erosion. It is rare, however, that a landscape becomes entirely barren, because soil that is “lost” by erosion is transported to other regions, bearing nutrients, organic matter and water. Such “bestowals” are prevalent particularly in South Asia, where “sediment harvesting” is possible, e.g. the nullah plugs in India. Hence, the “cause and effect” paradigm of soil erosion and crop productivity should be treated with caution, since an adverse effect on one location may transfer an advantage to somewhere else43.

Some confidence is justified in connecting soil erosion and crop yields, primarily on the basis of experimental runoff plots, where measured soil losses are related to both current and future yields, though not exclusively to the underlying mechanisms of soil erosion. As a general trend, plots of crop yield versus cumulative soil depletion (t ha-1) reveal curvilinear, inverse-exponential type relationships, i.e. the yield drops as the soil gets thinner. Hence, there is an initially sharp loss of productivity, followed by stages in which the impact is successively less. While alternative behaviour has been identified, the overall message is that it is comparatively easy to bring back slightly degraded land into economic use, and that the net returns are always better if the yield has not fallen to under 50%. In contrast, when land has been severely degraded, to bring it back into useful (or even economic) production is a tremendously difficult task. Some soils (Figure 7) are far more resilient than others, e.g. a Nitosol (clay-rich, and common on basic rocks in highlands) which may be orders of magnitude more resistant to productivity loss than an extremely sensitive soil, such as a Ferralsol or Acrisol (common in arid, humid tropical rainforests), especially under good levels of management43. However, it has been pointed out that the results reported from studies of the dependency of crop productivity on the degree of soil erosion, are inconsistent in respect to both the magnitude of the response and indeed the shape of the response curve. Accordingly, an analysis was undertaken44 to determine whether general patterns, or key features with either a physical or methodological origin, in terms of the measurement of soil loss, might be identified.

It was concluded that the experimental methodology employed has an overwhelming influence on the apparent magnitude of the soil loss: i.e. from a comparative-plot method, an average loss in crop productivity of 4.3% per 10 cm of soil loss was obtained; in contrast, results based on the transect method gave a 10.9% average; while those from measurements of desurfacing averaged at 26.6%. Although there was no identifiable effect of physical variables (water deficit, physical root hindrance, nutrient deficit) on the magnitude of the response curve, nor of the particular experimental method on its shape, it was found that water deficit and physical root hindrance caused convex curves, while nutrient deficit gave rise to linear or concave curves. As a general rule, it is only when the regressor (meaning an imbalance between the crop demand and soil supply) is nutrient deficit and the experimental method is desurfacing, that the curve is concave. As an explanation for this, we may note that with a single act of desurfacing (artificially removing soil), it is the upper topsoil that is removed, and because the concentration of nutrients tends to increase toward the surface, this has a more pronounced impact on yields than when subsequent, lower lying soil layers are removed. The timescales over which natural erosion occurs are sufficiently long that the nutrients are continuously redistributed throughout the soil, and so comparable quantities of nutrients are removed layer by layer. It is also the case that where nutrient deficits are obviated by the application of fertilizer, the response curves tend to be of a convex form. This suggests that further erosion is likely to cause reductions in crop productivity of increasing severity.

[Fig 7]

When comparing results for soil loss experiments, the exact experimental methodology should be taken into account, since it is clear that practically all the variation in data from otherwise comparable settings can be explained in terms of the different types of measurement employed, e.g. the results from desurfacing measurements were as an average six-fold greater than those derived using the comparative plot approach. It appears reasonable that the more sensible estimates of soil loss are obtained from comparative plot experiments, while the other methods may lead to gross overstatements about the adversity of intensive mechanised (industrialised) agriculture on the effect of soil erosion on crop yields. The aforementioned differences in the shape of the response curve according to different physical influences, might be accounted for in terms of the behaviour of the regressor in the soil profile as a nexus with the crop response to the regressor. For example, because the concentration of nutrients decreases, but in a nonlinear fashion, as a function of soil depth, successive removal of soil has an attenuating effect on crop yields. In contrast, the availability of water decreases more or less linearly with soil depth, but how well crops grow is increasingly impaired as the availability of water is reduced. When roots extend such that they come into contact with the restricting layer (e.g. bedrock, or clay), growth is hindered, as intuition might suggest, but since the density of roots tends to increase towards the surface, the effect of soil loss is more pronounced when shallower (top)soils are present.

In the case of gradual, rather than accelerated erosion, and the soil is well nourished by effective management practices, the only significant regressors are physical hindrance and water deficit. The indicated reduction in crop yields of 4% per 10 cm of soil lost implies that those regions that are subject to moderate erosion rates (ca 10 t ha-1 year-1 – i.e. of the order of 1 mm depth of soil lost per year) will suffer an average decline in crop yield of 0.4% each decade. The latter figure might appear marginal, yet for regions in which the soil erosion is more severe, the reduction in productivity might be worsened by an order of magnitude or greater. Increasing restriction of root growth is an allying consequence of compounding loss of soil depth, and accords with a progressive fall in crop yields. It is those soils with growth restricting layers, such as clayey subsoils, pans or bedrock to which particular attention should be paid. In terms of land management, and even in regions where soil erosion has not, as yet, resulted in particular declines in crop yields, it is likely that the steady loss of soil will experience increasingly marked losses in the future. As a warning, the generally convex shape of the crop-yield/erosion response curves indicate that particular care should be given to those soils that are already eroded, but are currently still productive. The heavy application of fertilizer is all that maintains the yields of some highly degraded lands, which are likely to plummet when the soil loss exceeds a certain degree. No one should be deceived by crop productivity alone, and the underlying condition of the soil and land must be taken into account, and monitored closely.

In a study of the effect of soil erosion on crop yields in Europe45, it was assumed that the applications of fertilizer compensate for the nutrient loss caused by soil erosion, and so this is not a significant factor in affecting crop productivity. Since it has been shown that rooting space and water availability are the major limiting factors in determining crop yields in eroding soils, the two elements were combined to form a single variable, called SWAP (soil water available to plants) which is considered to be the most important in accounting for the effect of soil erosion on crop yields. SWAP is determined as the product of soil depth and volumetric water content for that depth, integrated down to the normal rooting depth for a specific crop, which for wheat grown in Europe, is considered to be 1.2 metres. In principle, SWAP is the quantity of water (mm) held between field capacity (5 kPa) and the permanent wilting point (1,500 kPa), which for a soil in which cereals are grown is the sum of two components: water held at low suctions (5-200 kPa) and 0-1.2 m depth, and that held at higher suctions (200-1,500 kPa) and 0-0.5 m depth. The soil erosion data were obtained from the PESERA (Pan European Soil Erosion Risk Assessment) model. Thus, the authors made a prediction of the likelihood of soil erosion causing reduced crop yields over the next 100 years, and concluded that it was unlikely to be a problem in the productive ecosystems of northern Europe. In southern Europe, due to a combination of severe soil erosion from ancient times and slow soil formation rates (typical in Mediterranean climates because of low rainfall and rapid loss of SOC), the soils tend to be stony and shallow, with low wheat yields. The model predicts that, in contrast with northern Europe, in the southern regions, erosion-induced reductions of probably a few percent points are likely during the next 100 years.45 At a national level, the effect is projected to be most severe in Greece, Portugal, Spain and Italy, although it is not thought there will be any real threat to the agricultural productivity of Europe with the next 100 years.

The lifetime of a civilization, however, far exceeds one century and over the millennial timescale, the effects of such levels of erosion could integrate to become major impacts. Even if erosion were to prove of no direct threat to European agriculture, its control remains desirable due to its associated harmful environmental influences. While productivity might be maintained through larger inputs of artificial fertilizers, to compensate for losses of nutrients by soil erosion, agricultural sustainability is affected negatively: increased costs incurred through the manufacture of the fertilizer, with associated carbon emissions, and the runoff containing fertilizer, pesticide and herbicide may damage terrestrial and aquatic ecosystems, many of which are in a sensitive ecological equilibrium. Land whose productivity has declined may be abandoned, and thus soil erosion may drive changes in land-use. To explore this prospect in more detail, Lesvos, in the western region of Greece, was chosen46, since it has undergone accelerated erosion on marginal soils over the last century, and indeed there have been significant changes in land-use there. In 1886, some 3211 ha were under cereals, of which 53% had been converted to land only used for extensive grazing (rangeland) by the middle of the 20th Century. However, in neighbouring regions, cereals had returned to some extent in former rangeland, which indicates local scale changes which rendered some land better and other land worse for growing cereals, over time.

On the basis of a statistical analysis, it can be concluded that the physical features of the landscape have been critical in determining the abandonment and later reallocation of land under cereals. Thus, land with high slope gradients, high erosion rates and shallow soils tended to be removed from cereal production, while new arable land used for growing cereals tends to be in those areas were the soils are deep, the erosion rates low, and the slope gradients shallow. The logistic fit suggests an attribution of abandonment to the direct impact of erosion (25%) + the erosion/soil depth component (36%) + the direct impact of slope (39%). In regard to the reallocation of land, the direct influence of slope was much smaller (17%), but the influence of slope via the erosion/soil depth component was 80%. A possible reason for this is that farmers believed that the marginal productivity of the land that they abandoned had more to do with the shallowness of the soils than to steep slopes. It would then follow that soil depth mattered more when choosing land for reallocation than slope gradient. It is concluded that soil erosion is a significant driver in land-use change, though due to confounding effects it did not emerge as a significant independent variable in the analysis, and that the cultivation of cereals in western Lesvos will probably be abandoned within a few years.

Another study is reported of the response of soil erosion and sediment export to land-use change in four agricultural regions of Europe47, which over the past decades has been driven mainly by the introduction of new technologies. Thus, through the introduction of mechanized machinery, synthetic fertilizers, herbicides, pesticides and new cultivars, an increase in land productivity of 400─500% has been achieved. As a result, intensification has been the case in those regions where these new technologies could be suitably implemented, while abandonment or de-intensification occurred (reduction in inputs) in those areas that were less suitable. From a simulation of the response of erosion to land-use change over the past 50 years, it was inferred that de-intensification of land-use in marginal agricultural areas resulted in a strong reduction in erosion and the transport of sediments to rivers. This reduction in erosion is frequently enhanced by the conversion of a type of land-use that worsens erosion to one that is less harsh, e.g. the conversion of arable land to forest, on steeper slopes. The innate soil fertility also plays a role, since it is arable land with sandy and clayey soils (suitable but less erodible) that tends to be abandoned earlier, while the more long-standing arable land is typically that on silty soils, which are suitable but erodible. The issue of soil fertility and land area is crucial, if one result of climate change is that populations may move toward more polar regions. As is clear from Figure 8, the relative land area decreases, especially toward the south, and in the direction of both poles the soil tends to be of the poorer kind.

[Fig 8]

10. Peak oil, peak gas and peak phosphorus.

The greatest adverse impact on our system of industrialised agriculture would be the loss of a cheap supply of crude oil14 (“peak oil”), and the fuels, pesticides and herbicides that are derived from it. Although there is a cornucopian counterargument that peak oil can be disregarded, on the grounds that there are vast quantities of “oil” in the earth, it ignores or confounds what the term actually means. Specifically, peak oil refers to the rate of production of crude oil, not the size of the total hydrocarbon body there may lie in the multifarious reservoirs of global geology. To use an analogy, it is the size of the tap not the tank that matters. Much of the “oil” that remains will be recovered only with a far lower energy return than conventional crude oil, and much of what is claimed may not prove worth recovering at all. The bulk of the world’s tally is present in oil shale, and is not petroleum but a solid, primordial material called kerogen, which must be “cracked” (thermally decomposed) by heating it to 500 oC, in order to produce a liquid form that resembles crude oil2,14. Unsurprisingly, this requires a high input of energy, and which is comparable to the amount of energy that would be recovered by burning the resulting “oil”. It has also been proposed that “peak phosphorus” can be expected at some time during the present century, based on various analyses of how much phosphate rock there is available and its likely recovery rates15. Since phosphate rock is mined using machinery powered by oil-refined fuels, the loss of a cheap supply of crude oil would impact on the production of this principal source of phosphorus fertilizers, in addition to its restrictive influence on running farm machinery.

It is the occurrence of peak oil, and peak natural gas (a source of hydrogen, and hence ammonia from which nitrogen fertilizers are made), with their attendant consequences, that may invalidate many predictions made about how agriculture might prevail (and all other human activities for that matter), for the next 100 years, or even the next 20 years, since we may have to grow food largely in the absence of their inputs. In which case, protecting the soil is paramount. Having continued access to phosphorus is critical, and it has been proposed that more of this element might be got from the soil48. Indeed, the amount of phosphorus in soil is by far in excess of the amount that is mobile and hence available to plant roots; the vast majority being present in the form of insoluble compounds2. Moreover, the amount of phosphorus fertilizer that is applied to crops is probably twice that actually necessary because of the tendency to ignore the longer-term effect of the residual phosphorus in soil. Due to an historical lack of application of phosphorus fertilizer in Africa, it has been estimated that it will be necessary to apply 30─50% more of it to the soils there, and probably for a period of 30─50 years, in order to regain pre-depleted levels of the element48. From another study, it was concluded that phosphorus can be used more efficiently by both improving the uptake of it (P-acquisition efficiency) and a greater productivity per unit of P taken up by plants (P-use efficiency). The growth of crop-plants that have overall lower P-concentrations, and the undertaking of further research into the associated plant genetics is stressed49.