Chapter 4 - Status of degradation. I. Erosion and fertility decline

Back to contents - Previous file - Next file

Water erosion
Wind erosion
Soil fertility decline

Water erosion

(Tables 5 and 6, Figures 2 and 3)

According to the GLASOD assessment, a total of 83 M ha is assessed as affected by water erosion in the region, or 25 % of the total area under crops and pasture. This is made up of 33 M ha with light erosion, 36 M ha moderate and 13 M ha strong erosion. The dry zone is most affected with 39% of the area under crops and pasture, compared with 18% for the humid zone.

The countries most seriously affected are in absolute area India and Iran, and relative to crops and pasture, Iran, Sri Lanka and Nepal. Examples where erosion has reached the severe degree, leading to abandonment of land, include parts of the hill areas of Sri Lanka (Stocking, 1992; Sri Lanka, Natural Resources, Energy and Science Authority, 1991, p.120), and the Pothwar Plateau of the Punjab region of Pakistan¹ (Nizami and Shafiq, 1990). For current erosion under inappropriate land use, there are many estimates in excess of 100 t/ha per year, including for parts of India, Nepal and Sri Lanka (e.g. Das et al., 1991; Stocking, 1992).

The map shows a clear relation to physiographic units. Most affected are the populated mountain regions of the Himalaya-Hindu Kush, the mountainous rim of Iran, and the areas of predominantly rainfed agriculture of the Deccan of India (with the Western Ghats most seriously affected) and Sri Lanka. Also affected are strips where the Gangetic river system has cut into terraces, whilst ravines are widespread along the rivers Jumna and Chambal.

Table 6 shows some estimates of areas affected by water erosion, giving the words used to define the areas stated. For India, the earlier estimates are in the range 69-127 M ha, which is 2-4 times the GLASOD estimate. The figure of 4 M ha under gullies or ravines has frequently been quoted, and is one third that of the GLASOD value for strong degradation. The estimate of Sehgal and Abrol (1992) is a new assessment by the National Bureau of Soil Survey and Land Use Planning "following the criteria and guidelines of the GLASOD methodology". The value is over twice the original GLASOD estimate. For Pakistan, the totals are of the same magnitude, 11.2 as compared with 7.2 M ha.

These comparisons illustrate what will be found repeatedly, that estimates of areas affected by land degradation show a wide range of values.

TABLE 5 - GLASOD assessment: areas affected by water erosion (Unit: 1000 ha)

TABLE 6 - Country estimates of areas affected by water erosion

In terms of total area affected, water erosion is the most serious problem of land degradation in the region. It is the only degradation type which is widely found both in the dry and humid zones.

As the basis for discussion in the remainder of this report, the GLASOD estimates for water erosion are accepted, whilst noting that for India, it is possible that they are 2-3 times higher.

Figure 2 - Water erosion severity (GLASOD estimate)

Figure 3 - Erosion and fertility decline: GLASOD assessment

Wind erosion

(Tables 7 and 8, Figures 3 and 4)

In the GLASOD estimate, a total of 59 M ha is assessed as affected by wind erosion in the region, Iying entirely within the dry zone. Within this zone, 48% of land under crops and pasture is affected. This is predominantly, 34 M ha, of moderate degree. It is very unevenly represented by countries, affecting 60% of agricultural land in Iran and 42% in Pakistan, whilst the dry region of India has the same total area affected, 11 M ha, as Pakistan.

The map illustrates this clear and expected localization in the dry belt stretching from central Iran to the Thar Desert of Pakistan and India. The irrigated belt of the Indus system cuts a swathe through the affected zone, with wind erosion occurring along the unirrigated belts between river systems.

The relatively low proportion of Afghanistan mapped as affected by wind erosion is surprising, although the high altitude and consequent longer evapotranspiration of its low-rainfall areas may be partly responsible. The national report to the UNCED conference stases, "desertification and erosion continue unabated" (Afghanistan, Ministry of Planning, 1992). This situation requires clarification when political conditions permit.

Table 8 shows country estimates. For India, one estimate is similar to the GLASOD total, the others three times higher. For Pakistan, the country estimate is about half that of GLASOD. However, a recent land use survey of the whole country includes the mapping units, "range land, non-degraded" and "range land, degraded"; by inspection, it appears that over 90%, possibly 95%, of range land is considered to be degraded (Asian Development Bank, 1992b).

TABLE 7 - GLASOD assessment: areas affected by wind erosion (Unit: 1000 ha)

TABLE 8 - Country estimates of areas affected by wind erosion

As the basis for discussion in the remainder of this report, the GLASOD estimates for wind erosion are accepted.

Figure 4 - Wind erosion severity (GLASOD estimate)

Soil fertility decline

(Tables 9 and 10, Figure 3)

The GLASOD estimate

GLASOD defines this form of degradation as "loss of nutrients and/or organic master. The GLASOD assessment shows 65% of agricultural land in Bangladesh and 61% in Sri Lanka affected by this type of degradation. No other areas are reported apart from three map units in India, described on the data sheets as having "heavy leaching with lateritic crust formation". However, a recent country analysis of the GLASOD results gives a much larger value of 26 200 ha (Sehgal and Abrol, 1992).

It is clear that there is a reporting bias here. The respondents for Bangladesh and Sri Lanka recognize this form of degradation as being widespread on cropland, both rainfed and irrigated, whilst those for other countries of the humid zone initially did not (but see below). Evidence of the existence of this form of degradation calls for discussion.

Evidence for soil fertility decline

Over the past 30 years there has been a large increase in fertilizer consumption in the region, associated with the introduction of high-yielding crop varieties. Bangladesh, India, Iran, Pakistan and Sri Lanka all now apply on average more than 70 kg/ha nutrients. This has been a major factor in the increase in crop yields over the period.

However, an inter-related set of soil fertility problems has been reported, directly or indirectly associated with fertilizer application. An early report is from 1981 (Bowonder, 1981) and evidence is accumulating. These problems are as follows.

Organic master depletion Crop residues are widely used as fuel and fodder, and not returned to the soil. This results in a decrease in soil organic master content. In Bangladesh, the average organic master (presumably of topsoils) is said to have declined by 50%, from 2% to 1 %, over the past 20 years (Bangladesh, 1992). For the Indian State of Harayana, soil test reports over 15 years show a decrease in soil carbon (Chaudhary and Aneja, 1991). Decreased organic master leads to:

• degradation of soil physical properties, including water holding capacity, as has developed in India (Indian Council of Agricultural Research, persona! communication);
• reduced nutrient retention capacity;
• longer release of nutrients, including micronutrients, from mineralization of organic master.

As a consequence of all these effects, there may be longer response to fertilizer.

A continuing negative soil nutrient balance. Removal of nutrients from the soil in crop harvest appears substantially to exceed inputs as natural replacement and fertilizers. Negative soil nutrient balances have been reported for all three major nutrients in Bangladesh and Nepal; for phosphorus and potassium in Sri Lanka, and a large deficit for potassium in Pakistan (FAO, 1986b). Nutrient depletion has been reported for each of the 15 agro-climatic regions of India (Biswas and Tewatia, 1991; Tandon, 1992, citing other sources). For India, a deficiency between nutrient removal and addition of 60 kg/ha per year, or 9 Mt for the whole country, has been estimated (Tandon, 1992).

TABLE 9 - GLASOD assessment: areas affected by soil fertility decline*. (Unit: 1000 ha)

* Described in GLASOD as "Loss of nutrients and/or organic matter".

TABLE 10 - Soil fertility decline: revised estimates (Unit: 1000 ha)

* Of which 2 200 are attributed to the dry region.

Imbalance in fertilizer application Fertilizer use in the region is dominated by nitrogen; N:P and N:K ratios are higher than in the other parts of the world. For example, the N:P:K ratio for India is 1.00: 0.33: 0.17 compared with 1.00: 0.52: 0.40 for the world (FAO data; Pradhan, 1992). This trend originated in the early years of the 'green revolution'. When fertilizers are first applied to a soil, a high response is frequently obtained from nitrogen. The improved crop growth depletes the soil of other nutrients; "In such systems, nitrogen is simply used as a shovel to mine the soil of other nutrients" (Tandon, 1992). Long-term experiments in India show depletion of soil P and K are higher for plots with N fertilizer, and depletion of K still higher with N+P fertilizer (Tandon, 1992). In Pakistan, use of nitrogen (mainly as urea) is still increasing, whereas use of phosphorus has levelled off in the lest 5 years, and very little potassium or micronutrient fertilizers are applied (Twyford, 1994).

FIGURE 5 - Pakistan: kilogrammes of wheat produced per kilogramme of nitrogen supplied as fertilizer (Twyford, 1994)

Secondary and micronutrient deficiencies An increasing incidence of sulphur and zinc deficiency is occurring in the region. Sulphur deficiency has been reported for India, Pakistan and Sri Lanka, and zinc deficiency for India and Pakistan (FAO/RAPA, 1992, p.65; Bowonder, 1981; Chaudhary and Aneja, 1991; Abrol, 1990). For Bangladesh, 3.9 M ha are reported deficient in sulphur and 1.75 M ha in zinc, including areas of continuous swamp rice cultivation (Bangladesh, 1992; Shaheed, 1992). Pakistan, because of its generally alkaline soils, is particularly liable to micronutrient deficiencies, which are being increasingly reported (Twyford, 1994).

Failure of increases in fertilizer use to be matched by increases in crop yield A levelling off, or plateau, in the crop yield increases which took place in the 1960s and 1970s is found in many countries of the region. The situation is clearly illustrated by data for Pakistan, where more or less linear increases in fertilizer nutrient use have not been equalled by rates of yield increase for wheat, rice and sugar cane (Figure 5). There may be several reasons for this serious effect, but a major contributory factor is undoubtedly decline in soil productivity (Chaudhary and Aneja, 1991).

Lower responses to fertilizers Long-term experiments in India have shown low or zero response to N fertilizer under severe P deficiency, and a low (and uneconomic) response to N-P-K fertilizer where there is zinc deficiency (Tandon, 1992). A striking exemple is a 33-year fertilizer experiment at Ranchi, Bihar; despite changes to improved varieties, wheat yields have declined substantially over the period with N. NP and NPK fertilization, whereas they have risen with farmyard manure (Goswami and Rattan, 1992).

Despite the reports cited above, a statement has recently been made with respect to Bangladesh that, "On present evidence, it is difficult to establish any significant trends in soil fertility. That is mainly because of the lack of long-term monitoring studies" (World Bank, 1991).

The existence of such a view highlights the urgent need for study of these problems. Two methods are available:

1. Long-term experiments These should be maintained or, where necessary, established at a limited number of representative sites in countries of the region. Difficulties are sometimes experienced in justifying funding for long-term experiments, but they are of immense value, and consideration should be given to international support for a network.
2. Soil monitoring This is the monitoring of changes in soil properties over time, on a statistically-based selection of sites on farmland. A high degree of standardization of analytical methods is essential. Soil monitoring should become a major element in the work of national soil survey organizations (Young, 1991).

The above evidence does not indicate the areal extent of soil fertility decline, other than that it is extensive in the region. It is the objective of this study, however, to obtain best estimates, and for this purpose, an adjustment will be made to what is considered a reporting bias in the GLASOD estimates. Given the large areas (60-65 % of agricultural land) reported as having nutrient deficiency in Bangladesh and Sri Lanka, and the existence of reports as outlined above, it is tentatively, and conservatively, estimated that an additional 20% of the agricultural land of both India and Pakistan are affected by soil fertility decline, at least to a light degree.

Revised country estimates Whilst soil fertility decline was shown for India only for a small area, as the above evidence has accumulated its greater extent has been accepted. A recent estimate gives 26.2 M ha as affected by loss of nutrients. There is no corresponding estimate for Pakistan, but evidence of the widespread occurrence of fertility decline is equally strong.

Consequently, as the basis for the rest of this report, the GLASOD estimates for soil fertility decline are revised for India and Pakistan, as in Table 10.