The Archive: Projecting the impact of climate change on the availability of barley in Kenya, 2013-2033

Prepared by Karin Viergever and Richard Tipper

Reviewed by Veronique Morel

Executive Summary

Diageo has invested substantially in developing local supply chain of barley, sorghum and cassava in Africa. This programme is particularly well-developed in Kenya where Diageo, through its subsidiary, East African Brewing Ltd, (EABL), has made special efforts to engage with smallholder farmers to promote improved production methods. However, Diageo is concerned that climate change and related factors such as water scarcity may affect the viability of crops grown in Kenya now and into the future. Notably, barley appears to have become less reliable in certain parts of East Africa in recent years.The purpose of this study is to provide Diageo with specific projections on the availability of barley in Kenya over the next 20 years, taking into account the impacts of climate change.

Since rainfall levels in Kenya are predicted to increase or remain the same under most climate change scenarios, albeit with more extremes, the main constraint expected in terms of growing conditions for barley will be temperature. Since 1980, mean temperature in Kenya has risen by 0.5°C, causing an associated rise in altitudinal crop boundaries for barley of approximately 90 m. In other words, the area suitable for growing barley has shifted up by 90 metres of elevation. This study estimates how continued rising temperatures will affect the areas suitable for barley growth in Kenya as a result of the following three climate change scenarios:

• Scenario 1: a slow change scenario representing half the historic rate of altitudinal shift (2.25m per year)
• Scenario 2: a steady change scenario with continuation of the historic rate of altitudinal shift (4.5m per year) .
• Scenario 3: a rapid change scenario with double the historic rate of altitudinal shift (9m per year)

Figure A: Reduction in land area suitable for barley growing resulting from three climate change scenarios

Analysis of the projected impact of rising temperatures on areas suitable for barley cropping in Kenya showed a general decline over time. In the short-term (by 2018), the areas suitable for barley cropping are estimated to decline 4% under the slow change scenario and 16% under the rapid change scenario. Over the next 10 years (by 2023), the decline is estimated to be between 8% and 30% and between 16% and 53% by 2033.

It should be noted that the amount of land deemed suitable for growing barley as a result of this analysis is far in excess of that currently in production in Kenya (over 900,000 hectares of suitable land versus 17,500 hectares planted by EABL suppliers in 2013-2014). Thus, the impacts of a decline in land currently used for barley production may be offset by moving production into unaffected areas in the future.

Figure B: Graph presents estimated change in barley production for each of three climate change scenarios. Beige columns represent barley production for each scenario based on average annual yield of 2 t/ha. Blue columns represents production at 3.5 t/ha.

Another potential mitigation is to increase crop yield. Figure B presents an estimate of the impacts of increasing yield on overall barley production from current average of 2 t/ha to 3.5 t/ha (EABL goal), if 100% of suitable land is utilised. In Figure B, beige columns represent production loss over time in each climate change scenario if yields remain at 2 t/ha. Blue columns represent production resulting from yield increase to 3.5 t/ha. It is important to note, however, that data on barley production for 2010-2014 shows that the barley yield in Kenya is quite variable from year-to-year and between sourcing areas. Currently, average annual yields within the country vary between 0.6 t/ha and 4.6 t/ha, with a country-wide average yield of 2 t/ha.

Areas most vulnerable to rising temperature are Lower Narok, the area between Nakuru and Molo, and west of Moiben. Areas least vulnerable are likely to be Olokurto, Oloropil, Sakutiek, East of Timau and the large tract of suitable land east of Nakuru and Ndabibi. Figures 10-12 in the report show more detail.

The rate of loss of suitability will not be steady and can be expected to proceed faster or slower during certain periods depending on the interaction between the El Niño Southern Oscillation (ENSO), the Pacific Decadal Oscillation (PDO) and the Indian Ocean Dipole (IOD). While these phenomena are difficult to predict in the long term they can be monitored and forecast with some accuracy a few months in advance. It is worth noting that we have recently been experiencing a negative PDO phase, so the rate of loss is currently at the low end of the range.

Based on the findings of this study, key recommendations are as follows:

• Incorporate a capacity for heat tolerance into current effort to breed higher-yielding barley varieties. Work by Al-Karaki, et al (2007) found a strong relationship between temperatures in the weeks immediately following germination and the eventual yield of barley. This critical phase of growth determines the effectiveness of root establishment – which has a big influence on the ability of plants to withstand water shortage or other stresses in the later part of the growing season. While there was an optimum temperature of around 15°C, there were considerable differences between varieties, which indicates that there is potential to breed for tolerance to different temperature ranges; and
• In addition to efforts to increase yield through better farming methods and/or improved seed varieties, EABL may also benefit from actively targeting specific land areas for transfer to barley production. The results of this study can be used to identify specific areas that are least vulnerable to decline in suitability. This level of analysis can be done via Ecometrica’s online software program which will be available free of charge to Diageo for a limited amount of time after publication of this report. In such an analysis, the locations of specific farms can be analysed directly against research findings, enabling the user to determine whether the farm is located at the margin and thus at risk for decline or located in an area expected to remain suitable over time.

1. Introduction

Africa represents one of Diageo’s largest emerging market in terms of net sales, with the company producing and selling a range of local beer brands including Guinness, Tusker, Senator, Serengeti Premium Lager and Bell.

This has led Diageo to invest in the promotion of sourcing suitable quality barley, sorghum and cassava in several African countries and of implementing a policy of favouring locally grown ingredients. The company has made special efforts to engage with smallholder farmers to promote improved production methods. However, Diageo is concerned that climate change and related factors such as water scarcity may affect the viability of crops grown in these geographies now and into the future. Notably, barley appears to have become less reliable in certain parts of East Africa in recent years. Prior to this study, no quantitative analysis of the impacts of climate change to agricultural materials of interest to Diageo has been conducted.

Information about the likely viability of crops such as barley over the next 20+ years in specific areas is commercially relevant as it will help to answer the following questions:

1. Policy: does climate change represent a technical barrier to Diageo’s policy to support local sourcing of ingredients for African beers? What is the extent and urgency of this barrier? What areas will be most affected?
2. Sourcing implications: should Diageo invest in alternative sourcing arrangements for some crops? What are the cost implications? Should the company engage in negotiations on import tariffs?
3. Product planning: should Diageo consider changing the composition of ingredients to some of its products? What ingredients can be sustainably sourced from within the region over the long term?
4. Farmer partnerships: Should Diageo change its recommendations and/or relationships with farmers? Over what timescale? In which areas? Making decisions in these areas without information on the likely impact of climate change on cropland suitability could expose Diageo to higher systemic costs, less efficient supply chain arrangements and poor relationships with producers.

This project is focused on barley in Kenya for the following reasons:

• Barley in the tropics is more vulnerable to climate change than other brewing crops, and has a more limited range of suitable growing areas
• Barley production in Kenya is strongly dependent on demand for brewing
• Barley is an important ingredient for premium beer products
• Kenya has relatively good agricultural data

The aim of this study is to provide Diageo with projections of availability of barley in Kenya, in terms of area suitable for cropping as well as production potential, over the next 20 years. The analysis will be based on expected changes to the suitability of agricultural land for the cultivation of barley in Kenya in 5-year intervals, between 2013 and 2033.

2. Background on barley

Based on global FAO statistics for 2000-2012, barley (Hordeum vulgare L.) is, after maize, rice and wheat, the fourth most produced grain on an approximate dry weight basis (FAOSTAT, 2014, see Fig 1). Global barley volume has remained largely stable at approximately 150 million tonnes/year between 2000 and 2012 (versus maize which has increased from 600M to just under 900M tonnes). Although it does not tolerate highly humid warm climates, it is an adaptable cereal that has a growing range that extends from the sub-Arctic to upland areas of the subtropics. This suggests a gene pool of different varieties for wide environmental adaptability and good stress resistance (Cattivelli et al., 2011).

Figure 1: Global grain production 2000-2012, showing that barley is the fourth most produced grain on an approximate dry weight basis (data from FAOSTAT, 2014).

Barley has good drought, cold, and salt tolerance and is generally produced in temperate and semiarid subtropical climates. Barley production occurs at higher latitudes and altitudes and closer to the limits of deserts than any other cereal crop, i.e. barley is generally produced under conditions of moderate water stress. Ideal growth conditions are well-drained loam soils, at moderate rainfall (400 – 800 mm) or under irrigation, and at moderate temperature regimes (15 – 30°C) (Jaetzold, 2006a, Ullrich, 2011). Environmental abiotic stress factors that can cause severe grain losses are often caused by high or low temperatures, drought, anaerobiosis (i.e. the absence of oxygen in the soil), and soil anomalies such as excess salt. Such abiotic stresses often occur simultaneously, for example, drought (i.e. water availability below that required for maximum crop yield) is often associated with the occurrence of high temperatures. Soil salinity and soil sodicity are common problems in arid and semiarid areas, therefore barley varieties grown in these marginal areas have to be tolerant to soil salinity (Cattivelli et al., 2011).

Successful cropping, and crop yield, additionally relies on a number of constraining factors such as the level of farmer inputs of cultivar, fertilizer, pesticides, and irrigation. Yield averages range globally from approximately 2 t/ha (e.g. Australia) to more than 7 t/ha (e.g. the United Kingdom) (Ullrich, 2011). FAOSTAT reports an average barley yield for 2000-2012 in Kenya of 3 t/ha (see Fig. 5 (FAOSTAT, 2014)), however, EABL records between 2010 and 2014 show an average of 2 t/ha. Typically, country yield averages, hectarage, and total production reflect relative growing conditions (mainly related to precipitation) and management technology (mainly soil fertility and pest management) (Ullrich, 2011).

Several morphological and commercial forms of barley exist, including winter, spring, two – row, six – row, awned, awnless, hooded, covered, naked, hulless, and malting, feed (grain and forage), and food types (Ullrich, 2011). Globally, barley is predominantly used for animal feed, malt and food for human consumption. Thirty percent of global barley production is used for malting, 90% of which is utilized in beer making and the rest for distilling and food applications (Ullrich, 2011).

Brewing from barley has developed over thousands of years and has therefore brought about selection of barley for improvements in malting and brewing qualities. Soft and hulled barley are typically preferred for malting and brewing. Mainly six- and two-row are used in brewing, with different parameters (kernel size fraction, kernel weight, nitrogen content), specific to the end brew, determining the strain that is used (Gupta et al., 2010). In addition to the strain, good quality malt requires a high moisture content at harvest. In some areas, climate change is affecting the ability of producers to meet this minimum moisture level. For the breeding of malting barley, industry considers a range of traits relating to the germination process and the physical and chemical composition of barley and malt. However, actual malting of grain and wet chemical analyses are still the principal procedures for analysis and selection (Ullrich, 2011).

The average annual estimated global production of barley-based beer has rapidly increased between 2000 and 2012:

• Global: increase of 139% (136.7 vs 189.9 M t),
• Africa: increase of202% (6.2 vs 12.5 M t).
• East Africa: increase of 286.4% (1.1 vs 3.2 M t) (FAOSTAT, 2014).

3. Barley cropping in Kenya

3.1 Areas suitable for barley growing

Kenya has approximately 576,000 km² landmass, of which only about 16% (~92,160 km² ) is of high and medium agricultural potential with adequate and reliable rainfall. This potentially arable land is dominated by commercial agriculture with cropland occupying 31%, grazing land accounting for 30%, and forests occupying 22% (Government of Kenya, 2009).

The main conditions that affect suitability for barley are related to elevation and soil. Barley requires well-drained moderately fertile soils of medium texture and a pH of 6.5-8.0. Barley is intolerant to waterlogging, but tolerates salinity (Jaetzold et al., 2005). Elevation affects temperature and rainfall reliability and is therefore an important determining factor for crop suitability. Jaetzold et al. (2005) developed an agro-ecological sub-zonation for Kenya that identified bands of suitability for specific crops based on agro-ecological zones (see Fig. 2), further refined by models for the probability of meeting the temperature and water requirements¹ of each crop at various altitude belts. Areas most suitable for barley are those in the semi-humid and transitional zones.

Figure 2: Map showing agro-ecological zones for Kenya, as mapped by Sombroek et al. (1982), at 1: 1,000,000 (Data obtained from http://maps.virtualkenya.org/data). Jaetzold et al. (2005) based their agro-ecological zones on this work, supplemented by limiting factors for various crops. See 6.1.1 for more information.

Altitudinal limits could vary slightly depending on the location, e.g. the altitudinal limits in the West are generally approximately 200 m higher than in Central and East Kenya due to its location at the leeward side of the central highlands which receives the trade winds from the ocean, and due to the heating up of the large elevated land-mass of the Victoria Basin (Jaetzold et al., 2005).

Figure 3(a) shows areas suitable for barley growing based on location within the semi-humid and transitional agro-ecological zones (Fig. 2), combined with the altitudinal belts suitable for barley as defined by Jaetzold et al. (2005) and observed by Lawrence Maina (personal communication, 2014). Both sources agree that the lowest altitude at which barley is currently grown successfully in Kenya is 2,000 m. However, the lower altitudinal limit observed in some places is 1,700 m (L. Maina, personal communication, 2014), but specific locations have not been indicated. Jaetzold (2006a, 2006b) list altitudinal limits as low as 1,300 m for parts of Central and Eastern Kenya. Since these limits have not been confirmed in country, the analyses were carried out using lower limits of 2,000 m.

Figure 3(a): Map showing areas suitable for barley growing based on elevations above 2,000 m within the Semi-humid and Transitional agro-ecological zones (indicated by black lines on the map). The suitable areas in this map cover 1,213,332 ha. See 6.1.2 for information on the elevation data. It should be noted that this suitability map includes all slopes, see Fig. 4 for suitability taking into account slope.

Barley production data provided by EABL showed that, apart from the areas indicated in Fig 3(a) as suitable for barley production, barley is also produced in the region that includes Olokurto, Oloropil and Mau Narok. Since Olokurto and Oloropil, in particular, represent higher barley production areas, the barley suitability map was amended to include this region, shown in Fig 3(b).

Although barley tends to be better suited to higher altitudes, there is a fine balance between suitability and productivity. Rain tends to start earlier at higher altitudes, but growing periods lengthen at higher altitudes because the production of biomass is slower in cooler altitudinal climates. Therefore the chances of a crop ripening before the end of the rainy season becomes smaller in the higher belts (Jaetzold, 2005). This balance is affected when climate change causes temperature rises along the altitudinal belts.

Figure 3(b): Map showing areas suitable for barley growing based on elevations above 2,000 m within the Semi-humid and Transitional agro-ecological zones as shown in Fig 3(a) with the addition of current barley production areas at Olokurto, Oloropil and Mau Narok. The addition of these areas reflects the results of a validation exercise undertaken with Lawrence Maina to confirm sourcing locations. It should be noted that this suitability map includes all slopes, see Fig. 4 for suitability taking into account slope.

This study estimates the changes in the areas suitable for barley growth in Kenya due to the upwards movement of the lower limits caused by warming. Studies show that rainfall in Kenya is predicted to increase or remain the same (see section 4.1) and without further in-depth meteorological studies, it is difficult to predict how local rainfall (amounts and reliability) will be affected by climate change. For the purposes of this study we therefore assume that the agro-ecological zones stay unchanged and only take into account the effect of increasing temperatures. Since it is unlikely, due to the limitations of harvesting equipment such as combine harvesters, that barley will be effectively harvested on slopes greater than 18% (approximately 10 degrees), we have removed these from the area estimates of suitable barley growing areas, see Fig. 4. This represents a total suitable area of 936,068 hectares.

Figure 4: Map showing areas suitable for barley growing based on elevations above 2,000 m within the Semihumid and Transitional agro-ecological zones (indicated by black lines on the map), and with slopes of lower than 10 degrees. See 6.1.3 for information on the slope data.

3.2 Current barley growing areas and practices

Barley in Kenya is rainfed. All barley planted and bought by EABL in Kenya is malting barley. However, when harvested, only approximately 60% is of sufficient quality to be used as malting barley. Approximately 30% is used as adjunct barley (used in the brewing process with the addition of enzymes) while approximately 5-7% is used as feed barley.

Figure 5 shows annual production and yield for Kenya as recorded by FAOSTAT and annual production as recorded by EABL. Although the trends roughly correspond, there is a difference in the reported barley production from the two sources, with a notable difference in the data for 2001, with FAO data showing production at approximately 110,000 tonnes while EABL records indicate a number closer to 50,000 tonnes. The reason for the discrepancy between datasets is not known, however, given that EABL buys nearly 100% of all barley produced in Kenya, it is assumed that EABL data is the more accurate. The questionable peak in production for 2001 in FAO data also partly explains the higher average yield reported by FAO over the 20 year period 1992-2012. Again, EABL data on yields is assumed to be the more accurate and has been used throughout this research.

Figure 5: Barley production (in t, left vertical axis, green solid line) and yield (in t/ha, right vertical axis, blue line) for Kenya, 1992-2012 (data from FAOSTAT, 2014). Barley production statistics for Kenya 1997-2014 as provided by EABL (L. Maina, personal communication, 2014) are indicated in the green dashed line.

The most common malting barley varieties used in Kenya are Fanaka, Cocktail, Quench, Sabini and Nguzo. Although the latter two are currently being phased out as they are older varieties with lower yields, it is recommended that seed samples be preserved in the interest of retaining this elements of species biodiversity. Cocktail and Quench perform well at high altitude areas with a longer growth period, while the other varieties can be grown in different regions if the timing of planting is done well. The potential yield for many of these varieties is as high as 6 t/ha, but currently, in Kenya, average yield is approximately 2 t/ha² . EABL’s goal is to increase yield to 3.5 t/ha by 2017 (L. Maina, personal communication, 2014).

The barley growing season depends on the onset of rainfall, with the main rainfall season stretching from February to October³ . Ninety percent of barley is planted after the first rains of the season, with a small amount of barley in Timau planted “dry”. Most barley production areas in Kenya have a mid and late planting season, with only Timau and Kinangop having two planting seasons due to the availability of rain⁴ . Unreliable rainfall remains the biggest challenge to barley production, specifically drought after germination and excessive rainfall during harvesting (L. Maina, personal communication, 2014).

Most growers do not use a rotation system, and approximately 80% of producers leave the land fallow after a barley crop. Most of the barley growers use medium levels of farm inputs but many of the farmers don’t own their land and therefore they do not do much to enrich the soil (L. Maina, personal communication, 2014). Soil fertility depletion has been described as the major biophysical root cause of the declining per-capita food availability in smallholder farms in sub-Saharan Africa, with a decline from 150 to 130 kg per person over the past 35 years in overall crop production. Emerging evidence attributes this to insufficient nutrient inputs relative to exports, primarily through harvested products, leaching, gaseous losses and soil erosion. This results in yields that are about 2-5 times lower than the potential (Jaetzold et al., 2005). Farmers are only likely to adopt sound soil management if they are assured of return on their investment. Another factor that affects barley yields is poor crop management (L. Maina, personal communication, 2014).

Lawrence Maina provided information on current growing areas in Kenya that supply barley to EABL. The data was provided in the form of descriptions and data tables⁵ , which we have compiled and presented in map format. Figure 6 shows the barley growing regions in terms of the potential land area that is suitable for barley growing. This excludes land uses such as forests, settlements, and water that are not suitable for agricultural use. The maps also indicate whether EABL have classified the area as having either “High”, “Medium” or “Low” potential for barley growing; the main determining factor being the reliability of rainfall. The only two areas classified as having low suitability experience unreliable rainfall. The latter areas also receive low amounts of rainfall whereas all other “High” and “Medium” suitability areas receive either medium or high amounts of rainfall⁶ .

Figure 6: Map comparing the areas that literature shows are suitable for barley growing (shaded in green) with the potential areas (in ha) that are suitable for barley growth as classified by EABL (represented by circles that are proportional to the extent of potential area for barley growing). The potential area includes land that is suitable for barley cropping and excludes land uses such as forests, settlements, and water. The colour of the symbol indicates whether EABL have classified the area as having either “High”, “Medium” or “Low” potential for barley growing. Note that the statistics for Olchoro, Melili and Sakutiek were combined, indicated on this map at Sakutiek (see source data in see Appendix 2).

Annual data for the period 2010-2014 on actual areas planted, production and average yields per region are shown in Figure 7. These statistics were averaged over the five-year period and are depicted in map format in Figure 8. In Kenya, 60% of malting barley is grown within the area of Olokurto Oloropil, Melili, Sakutiek, Mau Narok and Olchoro. Large scale growers produce approximately 40% of the barley sourced by EABL. Moiben and Timau both have few but large scale growers supplying about 20% of the total output (L. Maina, personal communication, 2014). Comparison of Figs. 6 and 8(a) show that the average actual barley areas planted over the past 5 years are a fraction of the potential areas (less than 2%). The difference accounts for land used for a variety of other uses, including competing crops and land left fallow. Nakuru, Ndabibi and Moiben have been least utilised to their potential (on average < 40% of the potential area was used for barley over the past 5 years) whilst Oloropil (>90%) and Timau⁷ (60%) were the best utilised regions with large (>=4000 ha) potential barley areas and Molo (74%) and Maralal (63%) were the best utilised regions with small (<=1,500 ha) potential barley areas.

Figure 7 shows that Oloropil, Timau, Olokurto and Olchoro were the highest barley production areas for the period 2010-2014. Overall, the amount of barley produced was linked to the area planted, unless an exceptionally low or high yield caused an anomaly (e.g. Olokurto, 2012 and Timau, 2010). Average yield varied between years for each region, with a range of >1 t/ha yield over the past 5 year period not uncommon, even in the high production regions classified by EABL as High potential (e.g. Olokurto and Timau). The largest variation in average yield was observed in Naro Moro, which is classified by EABL as a Low potential area. It is also noteworthy that yields show no clear trend over this period for any of the regions. In conclusion, the large differences in yield between sourcing regions (as well as from year-to-year) indicates that a greater understanding is needed as to what causes these differences. Differences in local weather conditions and soil and crop management practicescan make a large-scale increase in yield particularly challenging.

Figure 7: Graphs showing observed (a) barley planted areas (in ha), (b) barley production (t) and (c) average barley yield (in t/ha), per region, for 2010-2014. Only areas with planted area >500 ha are shown. Data for Timau 1st and 2nd season were combined. The source data for these graphs are shown in Appendix;, these data are summarised in Appendix 7.

Figure 8: Maps showing averages, per region, of the (a) observed barley planted areas (in ha), (b) barley production (t) and (c) average yield (t/ha) for the period 2010-2014. Only areas with planted area >500 ha have been mapped, data for Timau 1st and 2nd season were combined. Data for “Upper Narok” was located at Narok and data for “Lower Narok” was located in the southern area of Narok county.

4. The effects of climate change on agricultural production

Climate change presents a major challenge for agriculture worldwide (IPCC, 2014). Research indicates that the most direct impact of climate change on crop yields will result from changes in temperature and rainfall (Beddington et al, 2011). Sub-Saharan Africa is expected to be particularly vulnerable to the impacts of higher temperatures and more erratic rainfall, because negative impacts are likely to be stronger in warmer regions where increases in temperature will have a larger impact on crop growth (Lobell et al., 2011), poorer farmers have less capital and resources to fall back on, and national support institutions in developing countries tend to be weaker. Continental Africa and Asia are expected to warm faster than the global average.

Despite a bleak outlook, farmers are consummate adaptors, who already cope with changes in market prices and variable weather on a seasonal basis. Farming systems have always changed over time, and there is evidence that farmers are already adapting to climate change by modifying their crop choices (Niggol and Mendelsohn, 2008).

4.1 Climate change impacts on land suitability in East Africa

The FAO defines land suitability classes for a given crop in terms of the probability of achieving the maximum yield, with Highly Suitable (S1) areas achieving, on average >80% of the maximum yield, Suitable (S2) achieving on average 60% to 80% of maximum yields and Moderate (S3) achieving 40% to 60%. Below this are grades of marginal or unsuitable land that provide progressively lower yields (FAO, 1996). Suitability for a specific crop is determined by a combination of soil, slope and climate conditions.

The current suitability ranges for existing production areas can be approximated by looking at the variations in barley yield (Figure 7c). The maximum potential yield for barley in Kenya appears to be around 4.5 tonnes per hectare. This was achieved in only 1 location (Naro Moro) in 1 year for which data was available. Most of the production areas produce on average 40% to 60% of the maximum yield, while several areas (Lower Narok, Maralal, Melili, Narok and Sakutiek) consistently produce under 2 tonnes per hectare, which puts them at the edge of moderate to marginal suitability. The average yield for Kenya between 2010 and 2014 was calculated to be 2 t/ha (see data in Appendix 7). All barley production areas experienced considerable variability in production, with 2010 being a particularly poor year for many areas. Since soil and slope are constant, we assume that most of this variability relates to climate and pests. Some variation may also be due to farming practices, for example, low yields could result from poor quality seed or reduced fertiliser inputs in some years.

Regional climate change models run by the North Carolina State University (Anyah et al, 2006) for East Africa indicate that:

• Average annual temperature will rise by between 1°C and 5°C, typically 1°C by 2020s and 4°C by 2100.
• Climate is likely to become wetter in both rainy seasons, but particularly in the Short Rain season (October to December). Global Climate Models predict increases in northern Kenya (rainfall increases by 40% by the end of the century), whilst a regional model suggests that there may be greater rainfall in the West.
• The rainfall seasonality i.e. Short and Long Rains are likely to remain the same.
• Rainfall events during the wet seasons will become more extreme by 2100, consequently flood events are likely to increase in frequency and severity.
• Droughts are likely to occur with similar frequency as at present, but to increase in severity.

Since rainfall in Kenya is predicted to increase or remain similar, albeit with more extremes, the biggest constraint expected in terms of growing conditions for barley will be the temperature. Excessive temperatures during the growing season reduce the rate of accumulation of starch within the growing barley grains. When combined with lower than average rainfall this has a strongly negative impact on yield, and when combined with excessive rainfall this can increase fungal diseases that reduce grain quality (Savin and Nicholas, 1996).

4.2 Uneven Rate of Change

It is important to recognise that climate change will not be experienced as a steady process, with each year measurably warmer than the previous. Underlying variability in the form of ocean atmosphere cycles of different, somewhat irregular, time periods will combine to accelerate the changes in some years and damp down or even reverse changes in other years.

Three important cycles that impact on the climate of East Africa, and will affect how change is felt, are:

1. The Pacific Decadal Oscillation (PDO) varies the rate at which heat is transported by deep ocean currents in the Pacific Ocean towards the poles from the equator. The PDO switches between positive and negative phases over periods of between 20 to 40 years. When the PDO is in a negative phase (as it has been since 2000) it extracts heat from the atmosphere into the ocean, reducing the rate of atmospheric warming around the world. When the PDO enters a positive phase we can expect the rate of change to accelerate as heat is released back to the atmosphere (Met Office, 2013).
2. The El Niño Southern Oscillation (ENSO) is a shift in the distribution of surface temperatures in the Pacific Ocean, with anomalies repeating on a cycle of approximately 5 to 7 years. Anomalies typically last from 9 months to 2 years. El Niño anomalies (warming of the surface of the Pacific along the equator) can result in increased rainfall and cooler than average temperatures in East Africa, whereas the opposite La Niña (cooling of Pacific sea surface) can lead to hot dry conditions in East Africa.
3. The Indian Ocean Dipole (IOD), is an irregular shift in surface temperatures of the Indian Ocean that typically occurs once or twice per decade. Positive IODs result in abnormal cooling of surface sea temperature in the south eastern equatorial Indian Ocean and corresponding warming in the western Indian Ocean, bringing heavy rainfall over the east Africa and severe droughts/forest fires over the Indonesian region. Negative IODs can have the opposite effect.

The combined effect of these cycles coupled with other random variations means that the rate of change will not be steady, and will proceed faster in some years than others.

It is worth noting that at the time of writing there is a forecast of a strong El Niño event to develop over the course of 2014.

4.3 Water Footprint of Barley

There has been growing interest in the corporate social responsibility (CSR) community in understanding the “water footprint” of products and services. Water usage at breweries has already been studied extensively by Diageo, but the question of water requirements of growing barley has received less attention.

For barley the water requirement or evapotranspiration (sum of evaporation and transpiration) is approximately 10% more than a pasture grass during the growing season (120 to 150 days), and depending on the temperature during the growing season will be 450 mm to 650 mm of rain per season (FAO, 1986). Growing barley is therefore unlikely to affect the availability of water for other uses. Thus, changing from barley to other field crops will have a minimal impact on water availability (unless irrigation is used). Because the water consumption of rainfed barley is similar to grass, converting from barley to natural grasses or woodland is unlikely to have a beneficial impact on water availability downstream. Similarly, converting from barley to maize is unlikely to have a significant impact downstream, since maize has only slightly higher water requirements than barley.

5. Scenarios and assumptions

This study focuses on the effect that temperature rise could have on the extent of barley production areas in Kenya due to an upward shift in the lower boundary of the altitudinal band that indicates suitability for barley.

Jaetzold et al. (2005) reported an increase in the annual mean temperature in Kenya since 1980, with a mean temperature rise of 0.5°C during 20 years and a concomitant rise in altitudinal crop boundaries or limits of approximately 90 m.

This study will therefore compare three climate change scenarios:

• Scenario 1: a slow change scenario representing half the historic rate of altitudinal shift (2.25m per year).
• Scenario 2: a steady change scenario with continuation of the historic rate of altitudinal shift (4.5m per year).
• Scenario 3: a rapid change scenario with double the historic rate of altitudinal shift (9m per year).

The analyses in this study are based on the following assumptions:

1. We assume that rainfall is not affected. This can be included in a similar study later, but needs an in-depth meteorological study on how the amount and reliability of rainfall will be affected regionally and/or locally.
2. We assume that the temperature change, and the concomitant upward shift in the lower boundaries for barley suitability have a linear rate of increase. Three scenarios are included to indicate a range of different outcomes.
3. In the absence of data or information on the existence of upper limits in the altitudinal range suitable for barley growing, we assume that there will be no shift in the upper altitudinal band and that only the lower boundary will shift upward. That means that the results will only show a decrease in area. However, certain areas in the upper altitudes may become more suitable than they are currently and therefore have increasing yield, since rising temperatures are likely to extend the growing season.
4. For the purposes of this study it is assumed that the productivity/yield of barley is uniform in all areas, i.e. changes in yield due to the improvement or decline in growing conditions are not taken into account. As shown in Fig. 7, barley yields vary between regions and over time.
5. This study does not take into account feedback mechanisms triggered by temperature rise and indirect effects of climate change such as pests and disease.
6. It is assumed that all areas that have been shown by literature research and expert knowledge to be suitable for barley production (see Fig. 4) are in fact suitable. If expert review of the report indicates that specific areas should be excluded or added because of other specific local conditions, analyses can be adapted to take this into account.

6. Data and methods

6.1 Data

This section describes the data used. Each dataset is briefly summarised and the data sources are given, along with an explanation as to why the data was used. Where applicable, a reference is given to a figure in the report that shows the dataset.

6.1.1 Kenya agro-ecological zones

Figure 2 shows agro-ecological zones for Kenya, as mapped by Sombroek et al. (1982), at 1: 1,000,000. Jaetzold et al. (2005) based their crop suitability map on this dataset, supplemented by limiting factors for various crops, such as minimum rainfall, and based on model outcomes. Since the Jaetzold dataset was not available in digital format, and only covered parts of Kenya in hardcopy, we used the original Sombroek dataset for this study. Visual comparison of the barley growing zones on the paper copies of the Jaetzold maps showed very good comparison for the barley growing zones in Kenya based on agro-ecological zones combined with altitude (see Fig. 3). Source: http://maps.virtualkenya.org/data/geonode:agro_ecological_zones (Accessed 27 Feb 2014)

6.1.2 Elevation data

Elevation data is core to the study. The elevation data used in this project is from the Shuttle Radar Topography Mission (SRTM). NASA and the US National Geospatial-Intelligence Agency (NGA) acquired radar data from the shuttle Endeavour in February 2000. The data was then used to create the first near-global set of land elevations. The resolution is 3 arc-second (approximately 90 m). The data for Kenya was obtained free of charge from the USGS as 90 separate 1 degree tiles which were then mosaicked together. Elevation data is shown as background in order to include context in Figs. 4, 6 and 8. Source: https://lta.cr.usgs.gov/get_data (Accessed 24 Feb 2014)

6.1.3 Slope data

Slope data was used to exclude areas with a slope greater than 10 degrees (18%) because of the limitations of combine harvesters. GIS software was used to derive a slope map from the elevation data described in 6.1.2. The resulting slope map is shown in Appendix 4.

6.1.4 Kenya settlements

For visualisation and context, the coordinates of settlements and places of note used for the data analyses and indicated on maps in the report were obtained from 2 sources:

• Center for International Earth Science Information Network – CIESIN – Columbia University, International Food Policy Research Institute – IFPRI, The World Bank, and Centro Internacional de Agricultura Tropical – CIAT. 2011. Global Rural-Urban Mapping Project, Version 1 (GRUMPv1): Settlement Points. Palisades, NY: NASA Socioeconomic Data and Applications Center (SEDAC). Source: http://sedac.ciesin.columbia.edu/data/set/grump-v1-settlement-points (Accessed 27 Feb 2014).
• An additional data source used was Joint Research Centre Fuzzy Gazzetteer. (Source: http://dma.jrc.it/services/fuzzyg/ (Accessed 10 April 2014).

6.1.5 Kenya administrative boundaries

The boundaries for Divisions and Counties were obtained from Virtual Kenya. Source: http://maps.virtualkenya.org/search (Accessed 24 February 2014).

6.1.6 Kenya forest cover

Forest cover in Kenya is diverse and highly fragmented, and there was no obvious, high quality, recent forest map available. Official data sets are somewhat conflicting. The FAO Forestry Department (2010) reports forest area estimates for Kenya from various sources that range between 3 million and 17 million ha. According to Peltorinne (2004), natural and plantation forest covers approximately 3.4% (200,800 ha) of the total land area and approximately 15% of high potential land.

FAO’s Africover dataset (2000) portrays forest types in Kenya mapped at a 1:200,000 scale. The Africover dataset for Kenya was based on satellite data (Landsat) from 1995 and could be deemed too old to be used for the purposes of excluding forest areas in this study. However, more recent data in map format is not currently available. Since Africover was based on intensive interpretation of medium resolution satellite imagery, we have decided to use this dataset to represent forest cover. The results section will show where estimated areas of future barley land conflicts with the existence of forest cover as identified by the Africover forest data. We recommend that detailed field investigation is done to verify these results, should Diageo consider extending barley cropping into such areas.

One of the main reasons for the discrepancies in forest area estimates, is that different definitions for “forest” are used, these are linked to a minimum contiguous area (patch size), height at maturity and percentage crown cover. Since it could be argued that crops can be grown in wooded land with a sparse crown cover, we used a subset of the Africover dataset to identify and isolate forest types with a canopy cover larger than 40% for the purposes of this study.The following forest classes were included: “Closed trees”, “Closed trees on temporarily flooded land”, “Open trees (65-40% crown cover)”, “Multilayerd trees (broadleaved evergreen)” and “Mangrove trees”. The dataset is shown in is shown in Appendix 5. Source: www.wri.org/resources/data-sets/kenya-gis-data (Accessed 24 February 2014).

6.1.7 Kenya protected areas

Protected areas are estimated to cover 7,194 ha, i.e. 12% of Kenya’s total land area (Government of Kenya, 2009). A dataset that contains national parks and reserves as well as forest reserves and game sanctuaries was downloaded from Virtual Kenya. The national parks and national reserves were extracted from this dataset for use in this project. There was an offset in the geolocation of the dataset, which was corrected in a GIS using a National Geographic-style base map containing boundaries for National reserves as reference. At the time of writing this report, the Virtual Kenya site was under construction and additional information on the original source of the data could not be accessed. The dataset is shown in is shown in Appendix 5. Source: maps.virtualkenya.org/search (Accessed 24 February 2014)

6.1.8 Wetland areas

Wetland areas in Kenya are home to variety of biodiversity and are diverse in type and distribution and cover 2-3% of the country’s surface area. Some of Kenya’s wetlands, such as Lake Naivasha and Lake Nakuru, have been included in the Ramsar list of Wetlands of International Importance (Government of Kenya, 2009). Wetland areas were derived from FAO’s Africover dataset, obtained from WRI. The dataset is shown in is shown in Appendix 5. Source: www.wri.org/resources/data-sets/kenya-gis-data (Accessed 24 February 2014).

6.2 Methods

After the search for relevant spatial data (see section 6.1) and subsequent acquisition of the data, checks were carried out and corrections were made where needed. Suitable areas for barley growing were derived by combining the semi-humid and transitional agro-ecological zones and altitudes higher than 2000 m above sea level, see Fig. 3. The barley suitability areas were then further refined by excluding slopes steeper than 18% (approximately 10 degrees), see Fig. 4. The latter was used as the baseline for area estimates of barley suitability in 2013. Changes in the elevation of the lower altitudinal boundary were calculated for the relevant time intervals, based on the 3 climate change scenarios described in section 5. Calculated altitudes were rounded off (see Table 1). GIS software was used to derive new barley suitability areas by adjusting the lower altitudinal boundary as indicated in Table 1. Finally, the areas derived were adjusted to exclude existing forest, wetland and protected land. Total areas were extracted from the data, tabulated and presented in graph format, see section 7. The output maps showing barley suitability under the different scenarios for the different years were uploaded to Ecometrica’s web-mapping platform, Our Ecosystem⁸

Table 1: Projected new lower altitudinal boundary (in meters) in 5 year intervals, for 3 climate change scenarios.

7. Results

Figure 9(a) shows the estimated changes in suitable barley areas for 2018, 2023, 2028 and 2033 under the 3 different scenarios of temperature rise. The data is given in Table 2. The maps in Figures 10-12 show how the spatial distribution of the suitable areas is estimated to change over time for each scenario. Maps showing the estimated changes for each scenario, per year, are included in Appendix 6.

Analyses of the projected impact of rising temperatures on areas suitable for barley cropping in Kenya showed a general decline over time. Since rising temperatures cause an upward movement of the lower altitudinal boundary of suitability bands, this is not unexpected.

Figure 9(a): Estimated extent of suitable barley areas for each of the three scenarios in 5-year intervals up to 2033.

If barley yields stay stable at 2 t/ha, the decrease in barley production will be directly linked to declining areas of barley cropping. Figure 9(b) shows the estimated percentage change in barley production over the three scenarios at an average yield of 2 t/ha and 3.5 t/ha. The magnitude of estimated change in area (and production at an average yield of 2 t/ha) within the first 5 years ranges from a 4% to an 16% decline over the three scenarios, escalating over the following 15 years to a decline of 16% to 53%. If EABL are successful in increasing the average barley yield to 3.5 t/ha by 2017, barley production in 2033, as a percentage of the 2013 production, is expected to increase only 22% under scenario 2, which uses historic changes in lower altitudinal boundary. Under scenario 3 (worst case), barley production is estimated to decrease by 18% in 2033 if yield is increased to 3.5 t/ha. See Table 2 for details.

Figure 9(b): Graph showing percentage decline in barley production for 3 scenarios at 2 different average barley yields.

As discussed in section 4 a decline, over time, in the extent of land suitable for barley growing will not cause a proportional decline in barley planting; a time lag is expected to occur during which farmers will plant barley with varying degrees of success.

Table 2: Estimated changes in the extent of barley growing areas in 5-year intervals up to 2033, for 3 climate change scenarios.

⁹Reflect areas suitable for barley that are not also forest, wetland and/or protected land and thus off-limits for barley growing. See Appendix 8 for a breakdown of all suitable areas including off-limit areas..

¹⁰The areas given in this row are based on the estimated overall decrease in suitable area as shown in the second row of this table. To see how specific locations will be affected, we recommend the use of Diageo’s OE webmapping interface, which can be accessed at http://diageo.ourecosystem.com.

¹¹This is the barley planted area in Kenya average for 2010-2014 (Maina, personal communication).

Figure 10: Map showing the estimated change in extent of areas suitable for barley growing under Scenario 1, i.e. an annual upward shift in altitudinal boundary of 2.25m. Since the extent of suitable area decreases for future years with an upward shift in the lower altitudinal boundary and no change in the upper boundary, the map indicates the extent for all years as overlapping areas, i.e. layers showing barley extent are “stacked” with the largest extent at the bottom (brown) and decreasing extents layered on top (orange, followed by green and pink). In other words, areas shown in brown are only suitable in 2018 and will no longer be suitable in 2023; areas shown in orange will be suitable in 2018 and 2023 but will no longer be suitable in 2028, and so on until 2033 where only the area coloured pink will be suitable. Areas expected to experience a greater amount of change over the next 20 years show wider bands in different colours (e.g. west of Moiben), while areas expected to experience little or no change have edges that show pink only (e.g. the area east of Moiben).

Figure 11: Map showing the estimated change in extent of areas suitable for barley growing under Scenario 2, i.e. an annual upward shift in altitudinal boundary of 4.5m.

Figure 12: Map showing the estimated change in extent of areas suitable for barley growing under Scenario 3, i.e. an annual upward shift in altitudinal boundary of 9m.

8. Conclusions

Suitable areas for barley production in Kenya are expected to decline by 8% to 30% over the next 10 years due to temperature change if no adaptive measures are taken. By 2033, suitable areas for barley production are expected to decline by 16% to 53%. The loss of suitability may accelerate beyond the mid 2020s and under a rapid climate change scenario we would expect a decline of more than 50% by the mid- 2030s.

The production areas most vulnerable to decline are Lower Narok, the area between Nakuru and Molo, and west of Moiben (Figures 10-12). The least vulnerable production areas are Olokurto, Oloropil, Sakutiek, East of Timau and the large tract of suitable land east of Nakuru and Ndabibi (see Figures 10-12).

The rate of loss of suitability will not be steady and can be expected to proceed faster or slower during certain periods depending on the interaction between the El Niño Southern Oscillation (ENSO), the Pacific Decadal Oscillation (PDO) and the Indian Ocean Dipole (IOD). While these phenomena are difficult to predict in the long term they can be monitored and forecast with some accuracy a few months in advance. We have recently been experiencing a negative PDO phase, so the rate of loss is currently at the low end of the range.

An estimate of total land area suitable for barley growing in Kenya, taking into account elevation, slope, absence of protected areas and applying climate change scenarios reveals that there is significantly more suitable land present than is currently in production for barley (over 900,000 hectares suitable versus 17,500 hectares in production during 2013/2014). Thus, the option exists, at least theoretically, to expand planted area in one region as a mitigation to declining area in another. From a water footprint standpoint, such a change in land use from grassland to barley or from another crop such as maize to barley is unlikely to have a significant impact on water availability ‘downstream’ as barley’s water needs are similar to grasses and somewhat less than maize. Thus, changing from barley to other field crops will have a minimal impact on water availability (unless irrigation is used).

There is also the potential to compensate for loss of land suitability through increased yields (plant breeding and improved cultivation methods), at least for 2 to 3 decades, although historically, yields across Kenya have shown no increase over time.

We have not identified new areas that could become suitable with rising temperatures. Migration of barley production uphill may be constrained by many local factors such as soil quality, the type of landholdings, natural ecosystems, and local changes in rainfall reliability. This would require field surveys and data collection over time.

9. Recommendations

Adaptive measures to mitigate the impact of temperature increases due to climate change should be considered by EABL, including the following:

• Incorporate a capacity for heat tolerance into current effort to breed higher-yielding barley varieties. Work by Al-Karaki, et al (2007) found a strong relationship between temperatures in the weeks immediately following germination and the eventual yield of barley. This critical phase of growth determines the effectiveness of root establishment – which has a big influence on the ability of plants to withstand water shortage or other stresses in the later part of the growing season. While there was an optimum temperature of around 15°C, there were considerable differences between varieties, which indicates that there is potential to breed for tolerance to different temperature ranges; and
• In addition to efforts to increase yield through better farming methods and/or improved seed varieties, EABL may also benefit from actively targeting specific land areas for transfer to barley production. The results of this study can be used to identify specific areas that are least vulnerable to declines in suitability. This level of analysis can be done via Ecometrica’s online software program which will be available free of charge to Diageo for a limited amount of time after publication of this report. In such an analysis, the locations of specific farms are mapped directly against research findings, enabling the user to determine whether the farm is located at the margin and thus at risk for decline or located in an area expected to remain suitable over time.

In additional to adaptive measures, Diageo could benefit from monitoring indicators of the rate of temperature change to understand and anticipate how impacts may occur from year to year. Key indicators that could provide warnings of rapid changes are:

• Global / Regional: (a) Shifts in the PDO from negative to positive (b) Strong El Nino or La Nina events and changes to the Indian Ocean Dipole;
• Local: (a) Temperatures following germination and at grain filling, (b) Rainfall anomalies and accumulated degree days during the growing season, (c) Vegetation growth anomalies using satellite monitoring of vegetation growth indices.

References

Al-Karaki, G.N., Al-Ajmi, A., and Othman, Y. (2007). Seed Germination and Early Root Growth of Three Barley Cultivars as Affected by Temperature and Water Stress. American-Eurasian Journal of Agricultural and Environmental Science, 2(2): 112-117. http://www.idosi.org/aejaes/jaes2%282%29/2.pdf (Last accessed 23 April 2014).

Anyah et al. (2006). http://weadapt.org/knowledge-base/national-adaptation-planning/climatechanges-in-east-africa (Last accessed 23 April 2014).

Beddington, J., Asaduzzaman, M., Fernandez, A., Clark, M., Guillou, M., Jahn, M., Erda, L., Mamo, T., Bo, N. Van, Nobre, C.A., Scholes, R., Sharma, R. and Wakhungu, J. (2011). Achieving food security in the face of climate change: Summary for policy makers from the Commission on Sustainable Agriculture and Climate Change. Copenhagen, Denmark: CCAFS

Cattivelli, L., Ceccarelli, S., Romagosa, I. and Stanca, M. 2011. Abiotic Stresses in Barley: Problems and Solutions. (pp 282-306). In: Ullrich, S.E. (ed.) Barley: Production, improvement, and uses. WileyBlackwell, Ames, IA, USA. 637 p

FAO Agricultural Zoning Guidelines (1996). http://www.fao.org/docrep/W2962E/W2962E00.htm (Last accessed 23 April 2014).

FAO Forestry Department. (2010). Global Forest Resources Assessment 2010 Country Report, Kenya. FRA 2010/107. Rome.

FAO (1986). Water and Irrigation Training Manuals. http://www.fao.org/docrep/S2022E/S2022E00.htm (Last accessed 23 April 2014).

FAOSTAT. (2014). Data obtained from FAOSTAT database, http://faostat3.fao.org/faostatgateway/go/to/download/Q/QC/E (Last accessed 12 February 2014).

Government of Kenya. (2009). Agricultural Sector Development Strategy 2009-2020. Nairobi.

Gupta, M., Abu-Ghannam, N., & Gallaghar, E. (2010). Barley for Brewing: Characteristic Changes during Malting, Brewing and Applications of its By-Products. Comprehensive Reviews in Food Science and Food Safety, 9, 318-328.

IPCC (2014) Food Security and Food Production Systems. WGII AR5 Chapter 7. Jaetzold, R., Schmidt, H., Hornetz, B., and Shisanya, C. (2005). Farm Management Handbook of Kenya Vol. II Part A, West Kenya. The Ministry of Agriculture, Nairobi, Kenya 2nd Edition.

Jaetzold, R., Schmidt, H., Hornetz, B., and Shisanya, C. (2006a). Farm Management Handbook of Kenya Vol. II Part B, Central Kenya. The Ministry of Agriculture, Nairobi, Kenya 2nd Edition.

Jaetzold, R., Schmidt, H., Hornetz, B., and Shisanya, C. (2006b). Farm Management Handbook of Kenya Vol. II Part C, East Kenya. The Ministry of Agriculture, Nairobi, Kenya 2nd Edition.

Lobell, D.B., Banziger, M., Magorokosho, C., Vivek, B. (2011). Nonlinear heat effects on African maize as evidenced by historical yield trials. Nat. Clim. Change, 1:42–45.

Met Office. (2013). The recent pause in global warming (2): What are the potential causes. www.metoffice.gov.uk/media/pdf/q/0/Paper2_recent_pause_in_global_warming.PDF (Last accessed 23 April 2014).

Niggol, S.S and Mendelsohn, R. (2008) An analysis of crop choice: Adapting to climate change in South American farms. Ecological Economics, 67(1):109-116.

Peltorinne, P. (2004). The forest types of Kenya. In: Pellikka, P., J. Ylhäisi & B. Clark (eds.) Taita Hills and Kenya, 2004 – seminar, reports and journal of a field excursion to Kenya. Expedition reports of the Department of Geography, University of Helsinki, Helsink,ISBN 952-10-2077-6, 148 pp.

Savin R. and Nicolas, M.E. (1996) Effects of Short Periods of Drought and High Temperature on Grain Growth and Starch Accumulation of Two Malting Barley Cultivars. Australian Journal of Plant Physiology 23 , 201–210.

Sombroek, W.G., Braun, H.M.H., van der Pauw, B.J.A. (1982). Exploratory Soil Map and Agro-climatic Zone Map of Kenya (Scale 1:1,000,000); Exploratory Soil Survey Report No E1; Kenya Soil Survey; Nairobi, Kenya, 1982.

Ullrich, S.E. (2011). Significance, adaptation, production, and trade of barley. (pp3-13) In: Ullrich, S.E. (ed.) Barley: Production, improvement, and uses. Wiley-Blackwell, Ames, IA, USA. 637 pp

Appendix 1 – Barley growing seasons per region

The table below was provided by Lawrence Maina of EABL. It contains information on the barley growing seasons for each of the regions. Note that only Timau and Kinangop have 2 growing season. However, as can be seen in the tables in Appendix 3, Kinangop is a low production area and was therefore not included in the discussions in Section 3.2.

Appendix 2 – Suitability and potential production areas of barley growing areas

The table below was provided by Lawrence Maina of EABL. It indicates potential production areas (this includes all agricultural land that is suitable for barley growing), as well as a summary of the amount and reliability of rainfall, a classification of the potential for barley growing and a description of soil type.

Appendix 3 – Yield performance data, 2010-2014 seasons

The tables below were provided by Lawrence Maina of EABL and contain barley production data, per region, for the 2010-2014 growing seasons. Averages of these data for area planted, production (delivery) and average yield are depicted in map form in Fig 8 in the report. A summary of areas planted (ha), production (t) and yield (t/ha) is given in Appendix 7.

Appendix 4 – Slope data

Figure A1: Map showing slopes derived from the SRTM digital elevation data

Appendix 5 – Forest, wetland and protected areas data

Figure A2: Map showing the location and extent of National Parks and Reserves, Wetland and Forest areas used during analyses. These areas were deemed unsuitable for barley growing. Background layer shows the digital elevation data, with lighter shades indicating higher elevations.

Appendix 6 – Maps showing estimated changes, per year, in the extent of barley suitability for each scenario

Figure A3: Map showing the estimated change in extent of areas suitable for barley growing in 2018 for each of the scenarios. Note that the map indicates the extent for all scenarios as overlapping areas, i.e. the scenario layers are “stacked” with the largest extent at the bottom and decreasing extents layered on top. This means that the extent of barley for scenario 1 is indicated by combination of the colours indicating barley extent for all scenarios.

Figure A4: Map showing the estimated change in extent of areas suitable for barley growing in 2023 for each of the scenarios.

Figure A5: Map showing the estimated change in extent of areas suitable for barley growing in 2028 for each of the scenarios

Figure A6: Map showing the estimated change in extent of areas suitable for barley growing in 2033 for each of the scenarios.

Appendix 7 – Data table accompanying Figure 7

The table below summarises areas planted (ha), production (t) and yield (t/ha) for the data tables shown in Appendix 3. It summarises the data that is depicted in Figure 7.

Appendix 8 – Areas suitable for barley that fall in forest, wetland and protected areas

This table gives a breakdown of areas that are suitable for barley but fall in forest, wetland and protected land and are therefore deemed off-limits for barley growing. This table accompanies Table 2 in the report.

¹²See Appendix 8 for a breakdown of areas that are suitable for barley but fall in forest, wetland and protected land and are therefore deemed unsuitable for barley growing.

Citations

¹The ratio of the average annual precipitation to the average annual evapotranspiration.

²The FAO data for Kenya suggests an average yield of 3 t/ha (see Fig 5), but the actual average yields observed vary between approximately 0.6 t/ha and 4.6 t/ha (see Fig. 7 and Appendix 7).

³A table showing planting and harvest dates for the different regions is included in Appendix 1.

⁴Kinangop was not included in Figures 6 and 7 as the area had planted areas < 500 ha.

⁵See Appendices 2-3 for the data tables provided.

⁶See Appendix 2 for the data table.

⁷It should be noted that this is based on the combined production in the 1st and 2nd growing season for Timau which, when separated out, will show that Timau is less effectively utilised.

⁸Diageo’s OE application can be accessed at http://diageo.ourecosystem.com. It is secure and is therefore only accessible by users with a login account. App administrators (Roberta Barbieri, Michael Wilson and Jos van Oostrum) can invite other users. For enquiries, please contact info@ecometrica.com