To date, the approach taken in the literature to describe chickpea environments has been somewhat fragmented. Chickpea habitats in WANA have been well described, with soil types, rainfall and temperature ranges mapped for each country individually, and more detailed information graphed for key locations (Saxenaet al., 1996). Outside the WANA region there is far less information available. Broad environmental categories (Khanna-Chopra and Sinha, 1987)
or key centres of cultivation (van der Maesen, 1972) have been represented by single locations and graphs of temperature, rainfall and photoperiod over the year presented. Although this provides a clear snapshot of gross differences, it tends to oversimplify reality and makes comparisons within and between regions difficult. Importantly in the context of this chapter, it does not encour- age an ecophysiological approach to investigating adaptation, because the finer points of habitat characterization are not captured. With the advent of high- resolution interpolated climate surfaces and user-friendly geographical infor- mation system (GIS) software freely available in the public domain (Hijmans et al., 2001, 2005; New et al., 2002), the task of characterizing chickpea habi- tats has become much simpler (Figs 3.1 and 3.2). Moreover, multivariate tech- niques such as principal components analysis can effectively integrate related variables and present a more holistic habitat characterization than mapping key descriptors individually.
Global chickpea distribution was defined by plotting all accessions with pass- port data from the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), Hyderabad, India; International Centre for Agricultural Research in the
−4
−3 -2
−1 0 1 2 3 4
−4 −3 −2 -1 0 1 2 3 4
−1.0
−0.5 0.0 0.5 1.0
−1.0 −0.5 0.0 0.5 1.0
WANA South Asia South America North America Europe East Africa Central Asia Australia Loadings
Altitude
Annual rainfall Summer rain proportion Winter temperature Rainfall CV
Summer temperature
Winter rain proportion
PC1: 37.1%
PC2: 35.4%
Fig. 3.1. Principal components analysis of climate of global chickpea growing regions.
Factors were quartimax rotated with Kaiser normalization to maximize the information summarized within single dimensions for PC1 and PC2, so that these scores could be mapped effectively (Fig. 3.2a and b).
Dry Areas (ICARDA), Aleppo, Syria; and United States Department of Agriculture (USDA) collections (n = 7759) to establish approximate boundaries within regions.
These were modified using pre-existing maps wherever possible (Cubero, 1980; van der Maesen, 1984; Saxena et al., 1996). This approach captures chickpea distribu- tion at a single time point, and may not be strictly accurate, as cropped areas fluctu- ate annually. However, it has the advantage of defining habitats for germplasm that has been collected, and is available for use by breeders and scientists. For regions without extensive germplasm collections or pre-existing maps (i.e. North America and Australia), growing areas were defined with the assistance of local chickpea breeders. Distribution shapefiles were converted into 10’ grids (~16 km resolu-
PC1 colour categories Increasing winter rain proportion.
Decreasing annual rainfall and summer rain proportion
Decreasing winter rain proportion.
Increasing annual rainfall and summer rain proportion
<−1.12
>=−1.12 < −0.67
>=−0.67 < −0.29
>=−0.29 < 0.15
>= 0.15 < 0.7
>= 0.7 < 1.15
>= 1.15
(a)
Fig. 3.2. (a) Global chickpea distribution classifi ed by PC1 scores capturing the
‘Mediterranean-ness’ of the climate (see Fig. 3.1 and Table 3.1). Navy blue zones are particularly Mediterranean, while red is the subtropical extreme. (b) Global chickpea
distribution classifi ed by PC2 scores capturing altitude (negative direction), rainfall variability, summer and winter temperatures (positive direction) (see Fig. 3.1 and Table 3.1). Altitude tends to decrease, while rainfall variability, summer and to a lesser extent winter temperature tend to increase as the chickpea growing region moves from navy blue to red. Note: while the Americas have been moved to the east to save space in the fi gure, latitudes have not been modifi ed.
PC2 colour categories Increasing altitude.
Decreasing rainfall variability, summer and winter temperature
Decreasing altitude.
Increasing rainfall variability, summer and winter temperature
<−1.14
>=−1.14 < −0.69
>=−0.69 < −0.16
>=−0.16 < 0.28
>= 0.28 < 0.71
>= 0.71 < 1.12
>= 1.12
(b)
tion) using DIVA-GIS (Hijmans et al., 2001) and climate data extracted (Hijmans et al., 2005) for each grid point. Subsequently all areas at an altitude >3000 m were excluded because a search of germplasm passport data suggested that chickpea was unlikely to grow at these elevations. In the northern hemisphere summer is from June to August and winter from December to February. While this summer period does not define the hottest annual quarter for Ethiopia and parts of South Asia, it does provide consistency across the globe, which facilitates interpretation of the results. In the southern hemisphere the seasons were reversed: winter is from June to August and summer from December to February. Multivariate trends in the data were analysed with SPSS v. 14.
Principal components analysis (PCA) demonstrated regional climatic simi- larities and differences very effectively, explaining almost 73% of total variance in two components (Fig. 3.1). PC1 captured the ‘Mediterranean-ness’ of the climate by positively loading annual precipitation and the proportion falling in summer, and negatively loading the proportion of rain falling in winter (Table 3.1). Thus, total rainfall and the proportion of summer rain tend to increase, while winter rainfall decreases from left to right in Fig. 3.1. In PC2 summer tem- perature was contrasted with altitude (Table 3.1). Rainfall variability and winter temperatures were also positively loaded on PC2, but their respective vectors run at almost 45° through the upper right quadrant of Fig. 3.1 because of their comparatively high PC1 coefficients. The chickpea regions largely clustered into two separate climate types:
1. Central Asia, WANA, Europe, much of South and North America and parts of Australia form a ‘Mediterranean-type’ group on the left side of Fig. 3.1, characterized by consistent, low annual rainfall, low winter temperatures and winter-dominant precipitation. Locations in the upper left quadrant of Fig. 3.1 represent stressful environments with a very narrow window for growth. These areas in the Egyptian Nile valley, central Iran, Iraq and Pakistan receive the lowest annual precipitation, predominantly in winter, and combine the highest summer temperatures with moderate winter temperatures.
2. South Asia, East Africa, parts of Australia and North and South America form a cluster largely on the right side of Fig. 3.1, and are characterized by high winter temperatures, high and variable rainfall, largely falling in summer. Note
Table 3.1. Factor loading for PCA of chickpea climate across the distribution range.
Trait PC1 PC2
Summer rain proportion (%) 0.84 0.42 Annual rainfall (mm) 0.82 0.01 Mean winter temperature (°C) 0.54 0.65 Monthly rainfall CV (%) 0.37 0.69 Mean summer temperature (°C) −0.04 0.96 Altitude (m) −0.12 −0.67 Winter rain proportion (%) −0.89 −0.19 Variance explained (%) 37.1 35.4
that both clusters have a wide spread of similar PC2 values, indicative of a wide range of summer temperatures and altitudes. Clustered around the origin is a large group of Australian locations characterized by an intermediate climate.
Mediterranean-type environments (low–medium PC1 scores)
Figure 3.2 gives a geographical context to the PCA. Chickpea throughout WANA, Central Asia, southern Europe, western USA, Chile and southern Australia is grown in typical Mediterranean climates with relatively low annual precipita- tion, which largely occurs in winter. Within these regions there are degrees of Mediterranean-ness, which are clearly highlighted in Fig. 3.2a and categorized in Table 3.2a. The navy blue zone in Fig. 3.2a is the most arid winter- dominant rainfall category, receiving an average of 219 mm rainfall/year, 51% of which falls in winter (Table 3.2a). These chickpea habitats are found in central Iran, central Pakistan, parts of Afghanistan, inner eastern Mediterranean, parts of North Africa, California, northern Chile and Western Australia. The navy blue zones in the southern hemisphere receive more annual rainfall than those in the north (Table 3.2a). In the Mediterranean basin chickpea is grown in the dark to medium blue zones, where annual rainfall is much higher (488–544 mm) and not as winter-dominant (31–43%, Table 3.2a). Moving north into Europe, chick- pea is found in the light blue to yellow zones with 640–700 mm annual rainfall (Fig. 3.2a), occurring equally in winter and summer (Table 3.2a). Similar trends are evident in other chickpea-growing regions. In the USA, California is con- siderably more Mediterranean than the Pacific North-west. In Chile there is a northerly trend of increasing Mediterranean-ness (Fig. 3.2a), where winter rain- fall proportion decreases from 66% to 50%, as annual rainfall increases from 330 to 1360 mm. In southern Australia there is a westerly trend of increasing Mediterranean-ness and a northerly trend towards summer- dominant rainfall along the east coast.
Summer-dominant rainfall environments (medium–high PC1 scores)
Most of the chickpea-growing regions in South Asia, East Africa and Peru are high-precipitation, summer-dominant rainfall environments (Fig. 3.2a). Within the Indian subcontinent there is a strongly decreasing rainfall cline from the south-east to the north-west (Table 3.2a), reflected in the colour gradient in Fig. 3.2a. Almost 50% of the chickpea production in India comes from Madhya Pradesh (Ali and Kumar, 2003), a central state enclosing much of the high rain and/or summer-dominant rainfall area in the subcontinent (dark red zone).
The light to dark blue areas in the north-west of the Indian subcontinent are not Mediterranean-type climates. These arid zones (<385 mm/year), largely in Pakistan, receive some winter rain (9–22%), but the bulk of annual rain falls in summer (48–61%, Table 3.2a). The blue zones in East Africa also largely reflect differences in annual rainfall (Table 3.2a). Only the navy blue area receives
53 Navy blue Dark blue Medium blue Light blue Yellow Orange Red
Region Annual rainfall
Australia 387 ± 6.4 370 ± 3.7 469 ± 3.8 604 ± 3 694 ± 4.8 – – Central Asia 245 ± 2.9 346 ± 4.4 433 ± 10.5 535 ± 15.5 552 ± 12.8 547 ± 5.8 – East Africa 197 ± 14.1 190 ± 7.4 326 ± 17.6 341 ± 16.5 542 ± 11.4 826 ± 5.3 1114 ± 8 Europe – 544 ± 3.9 553 ± 4.7 642 ± 6.5 764 ± 17.7 – – North America 241 ± 8 279 ± 19.1 349 ± 17.5 362 ± 2.1 405 ± 2.1 – – South America 330 ± 15.8 613 ± 23.1 975 ± 11.3 1212 ± 14.2 1361 ± 28.2 1167 2103 ± 207.9 South Asia 176 ± 10.4 147 ± 17.5 223 ± 5.8 385 ± 7.1 675 ± 3.7 988 ± 3.5 1635 ± 10.2 WANA 206 ± 3.1 488 ± 4.4 507 ± 4.4 558 ± 6.6 609 ± 88.5 – – Total 219 ± 2.6 475 ± 3.2 480 ± 3.4 523 ± 3.7 622 ± 3.4 959 ± 3.2 1468 ± 8.5
Winter rain (%)
Australia 50 ± 0.4 44 ± 0.3 27 ± 0.3 18 ± 0.1 12 ± 0.4 –
Central Asia 47 ± 0.5 36 ± 0.5 27 ± 0.9 19 ± 0.7 14 ± 0.2 11 ± 0.9 – East Africa 38 ± 0.9 30 ± 0.7 24 ± 1.5 11 ± 0.7 6 ± 0.3 4 ± 0.2 3 ± 0.1 Europe – 37 ± 0.2 31 ± 0.2 25 ± 0.2 24 ± 0.5 – – North America 51 ± 0.2 33 ± 0.7 27 ± 1 11 ± 0.2 8 ± 0.1 –
South America 66 ± 0.4 61 ± 0.3 56 ± 0.4 51 ± 0.3 50 ± 0.6 1013 ± 0.5
South Asia 39 ± 1.3 22 ± 0.6 14 ± 0.1 9 ± 0.2 3 ± 0.1 2 ± 0.03 2 ± 0.04 WANA 51 ± 0.2 43 ± 0.1 35 ± 0.2 30 ± 0.2 27 ± 2.9 –
Total 51 ± 0.2 42 ± 0.1 30 ± 0.2 19 ± 0.2 6 ± 0.1 3 ± 0.04 3 ± 0.04 Summer rain (%)
Australia 10 ± 0.3 12 ± 0.1 22 ± 0.3 36 ± 0.1 45 ± 0.7 – – Central Asia 1 ± 0.1 5 ± 0.4 12 ± 0.8 21 ± 0.9 27 ± 0.5 32 ± 1.7 – East Africa 26 ± 0.7 33 ± 0.6 38 ± 1.2 52 ± 1.5 42 ± 1.2 43 ± 0.7 60 ± 0.4 Europe –6 ± 0.1 13 ± 0.2 24 ± 0.3 24 ± 0.7 – –
North America 1 ± 0.1 19 ± 2 28 ± 1.9 43 ± 0.1 44 ± 0.1 – – South America 1 ± 0.1 2 ± 0.1 4 ± 0.1 6 ± 0.1 7 ± 0.3 38 36 ± 0.2 South Asia 23 ± 2 48 ± 1.4 58 ± 0.4 61 ± 0.5 57 ± 0.4 68 ± 0.1 64 ± 0.2 WANA 1 ± 0 4 ± 0.1 10 ± 0.1 13 ± 0.2 15 ± 1.3 – – Total 1 ± 0.1 6 ± 0.1 18 ± 0.3 34 ± 0.3 51 ± 0.3 64 ± 0.2 63 ± 0.2
more rain in winter than in summer. In the dark blue zone rainfall is equally distributed, and thereafter becomes increasingly summer-dominant (Table 3.2a).
The Mexican chickpea regions are arid (119–284 mm/year), summer- dominant rainfall areas, where the proportion of both summer and winter rain- fall increases (36–43% and 33–46%) with annual rainfall. In North America, moving eastwards from the Pacific North-west to the Northern Plains and Canada, rainfall increases to some extent (~360–410 mm/year), and becomes summer-dominant (43–44%).
Temperature and altitude (PC2)
Figure 3.2b puts PC2 into the geographical perspective. Altitude tends to decrease, while rainfall variability, summer and to a lesser extent winter temperature tend to increase as the chickpea-growing region moves from navy blue to red. Because winter temperatures and rainfall variability are not modelled by PC2 alone (Table 3.1), there are discrepancies in these variables between regions which are addressed individually below. From Europe to the Mediterranean there is a predominant south- erly trend of increasing PC2 scores punctuated by higher elevations in Spain, the Balkans, central Anatolia and eastwards into the Caucasus. From south-eastern Anatolia to northern Syria and Iraq, and from western to central Iran, the PC2 gradi- ent is particularly sharp (Fig. 3.2b), culminating in average summer temperatures of 32.5°C (Table 3.2b). Within PC2 categories mean winter temperatures in WANA and Europe are similar, ranging from −1.7°C to 12.9°C in WANA, and from 0.6°C to 10.5°C in Europe (Table 3.2b). However, altitudes and rainfall variability tend to be lower in any given category in Europe than in WANA (Table 3.2b).
South Asia, the Nile Valley, Mexico and subtropical Australia are dominated by high PC2 scores. In the Indian subcontinent there is a tight gradient in the north between the Terai and the Himalayan foothills, whereas in the plains PC2 scores tend to increase in a north-westerly direction in northern latitudes, and westerly in the south (Fig. 3.2b). These trends reflect the monsoonal pattern, which commences in June in the south, and a month later in the north and decreases average temperatures throughout. Throughout South Asia, peak annual temperatures are reached just prior to the onset of the monsoon. In fact, there is little difference in the mean temperature of the warmest quarter between the southern and northern halves of the subconti- nental chickpea distribution (30.8°C vs 31.9°C). However, mean winter temperatures differ markedly: 22.1°C in the south, compared with 16.8°C in the north. PC2 scores in Myanmar are generally the lowest in South Asia, but increase towards the warmer central zone. Winter temperatures and rainfall variability are underestimated by PC2 categories in South Asia, ranging from a relatively mild 12.5–20.6°C for the former to a comparatively high 90.7–138.4% for the latter (Table 3.2b). Chickpea in Central Asia is grown in areas with a wide range of temperatures and rainfall variability, tes- tament to the rapid changes in altitude found in the region. For Central Asia, winter temperatures are lower (−4.6–7.9°C) and altitudes higher (502–2101 m) than pre- dicted by the average values for the PC2 categories (Table 3.2b).
East Africa is dominated by low PC2 scores, indicative of high altitude and low mean summer temperature (Table 3.2b). Similar to South Asia, winter
55 Table 3.2b. Descriptive statistics for the principal chickpea-growing regions categorized by PC2 scores capturing altitude (negative
direction), rainfall variability, summer and winter temperatures (positive direction) (see Fig. 3.1 and Table 3.1).
Navy blue Dark blue Medium blue Light blue Yellow Orange Red
Region Altitude (m)
Australia – 612 ± 47 253 ± 7.3 213 ± 3.9 272 ± 3.8 126 ± 19.1 50 ± 4.7 Central Asia 2101 ± 39 1424 ± 55 923 ± 46 617 ± 30 615 ± 26 502 ± 34 924 ± 32 East Africa 2194 ± 14 1650 ± 11 1287 ± 10 1004 ± 17 792 ± 11 601 ± 10 299 ± 14 Europe 1013 ± 16 494 ± 12 452 ± 10 250 ± 7 96 ± 12 – – North America 810 ± 7 718 ± 7 322 ± 19 – 62 ± 8 246 ± 27 32 ± 6 South America 1222 ± 99 635 ± 67 414 ± 31 358 ± 21 189 ± 19 – – South Asia 1529 ± 43 1204 ± 12 826 ± 15 517 ± 11 409 ± 5 303 ± 3 181 ± 2 WANA 1710 ± 13 1465 ± 15 1274 ± 17 960 ± 18 755 ± 15 787 ± 11 482 ± 14 Total 1541 ± 11 1085 ± 10 910 ± 10 611 ± 9 524 ± 6 412 ± 5 218 ± 3
Mean summer temperature (°C)
Australia – 22.2 ± 0.1 23.2 ± 0.1 25.3 ± 0.1 26 ± 0.1 26.1 ± 0.1 31.2 ± 0.1 Central Asia 17 ± 0.2 21.7 ± 0.1 23.9 ± 0.1 25.4 ± 0.1 27.4 ± 0.1 29.3 ± 0.1 30.7 ± 0.1 East Africa 16.9 ± 0.1 20.3 ± 0.1 22.5 ± 0 24.4 ± 0.1 26.7 ± 0.1 28.3 ± 0.1 32.1 ± 0.1 Europe 17.9 ± 0.1 20.2 ± 0 22.7 ± 0 24 ± 0.1 25.8 ± 0.1 – – North America 18.1 ± 0.05 20 ± 0.02 21 ± 0.09 – 24.8 ± 0.13 26.7 ± 0.14 30.1 ± 0.13 South America 14.7 ± 0.5 17.5 ± 0.3 18.4 ± 0.1 19.7 ± 0.1 21.4 ± 0.3 – – South Asia 20.8 ± 0.18 22.5 ± 0.07 24.5 ± 0.07 25.9 ± 0.05 27.4 ± 0.03 29 ± 0.02 31.5 ± 0.03 WANA 18.7 ± 0.04 21.6 ± 0.03 24 ± 0.04 25.9 ± 0.05 27.6 ± 0.06 30.5 ± 0.06 32.5 ± 0.08 Total 18 ± 0.03 20.8 ± 0.02 23.2 ± 0.03 25.3 ± 0.03 27.3 ± 0.03 29.3 ± 0.02 31.6 ± 0.03
Mean winter temperature (°C)
Australia – 8.4 ± 0.3 9.7 ± 0.1 11.9 ± 0 12.8 ± 0.1 14.3 ± 0.2 24.3 ± 0.1 Central Asia −4.6 ± 0.2 −1.1 ± 0.3 1.3 ± 0.2 2.8 ± 0.1 3.9 ± 0.1 5.5 ± 0.1 7.9 ± 0.1 East Africa 16.5 ± 0.1 19.1 ± 0.1 20.7 ± 0.1 22.5 ± 0.2 24.2 ± 0.2 24.9 ± 0.2 24.6 ± 0.2
Continued
J.D. Berger and N.C. Turner Navy blue Dark blue Medium blue Light blue Yellow Orange Red
Region Mean winter temperature (°C)
Europe 0.6 ± 0.1 1.3 ± 0.2 6.5 ± 0.1 9.5 ± 0.1 10.5 ± 0.1 – – North America −10.3 ± 0.1 −8.4 ± 0.12 1.8 ± 0.08 – 8.5 ± 0.14 13.9 ± 0.67 18.8 ± 0.14 South America 5.4 ± 0.9 8.6 ± 0.7 8.4 ± 0.1 9.7 ± 0.1 11.5 ± 0.1 – – South Asia 12.5 ± 0.42 14.6 ± 0.17 17 ± 0.15 20.4 ± 0.12 20.6 ± 0.07 19.5 ± 0.06 17 ± 0.05 WANA −1.7 ± 0.06 1 ± 0.05 4 ± 0.07 7 ± 0.08 8.1 ± 0.08 8.6 ± 0.06 12.9 ± 0.11 Total 0.6 ± 0.16 2.3 ± 0.15 8.9 ± 0.12 11.7 ± 0.12 15.5 ± 0.12 17.1 ± 0.1 16.9 ± 0.05
Monthly rainfall CV (%)
Australia – 27.9 ± 3.4 26.1 ± 0.6 41.2 ± 0.5 68 ± 0.6 87.3 ± 1.5 119.8 ± 0.2 Central Asia 69.1 ± 1.8 73 ± 3 70.7 ± 2.5 75.2 ± 1.7 93.9 ± 1.2 103.9 ± 1.2 127.6 ± 2.7 East Africa 88 ± 0.9 95.2 ± 1.3 95.3 ± 1 108.8 ± 1.4 105.4 ± 1.7 107.2 ± 2.3 139.3 ± 5 Europe 29.1 ± 0.3 32.6 ± 0.4 44.2 ± 0.5 58.4 ± 0.4 66.3 ± 0.4 – – North America 61.9 ± 0.38 69.1 ± 0.32 42.1 ± 0.31 – 87.8 ± 0.52 103.1 ± 2.54 116.3 ± 1.12 South America 8 3 ± 3 82.2 ± 2.8 99.3 ± 1 110.7 ± 0.9 113.3 ± 1.4 – – South Asia 90.7 ± 1.3 89.6 ± 0.89 95.4 ± 0.79 103.7 ± 0.62 110.2 ± 0.48 128.1 ± 0.4 138.4 ± 0.42 WANA 56.8 ± 0.43 67 ± 0.55 75.8 ± 0.57 83.1 ± 0.51 88 ± 0.45 96.1 ± 0.45 114 ± 2.32 Total 60.7 ± 0.42 63.2 ± 0.46 69.8 ± 0.51 79.6 ± 0.48 99.5 ± 0.38 120.2 ± 0.4 135.2 ± 0.5
temperatures and rainfall variability are relatively high, ranging from 16.5°C to 24.9°C, and from 88.0% to 139.3%, respectively. As was the case with PC1, Ethiopia has a wide range of PC2 scores in close proximity, indicative of a wide habitat diversity, which is reflected in the diversity of the germplasm found there (Harlan, 1969).
In North America PC2 scores increase with latitude along the west coast, and from the Pacific North-west to the Northern Plains (Fig. 3.2b). Medium to navy blue zones in the north are particularly cold in winter, ranging from
−10.3°C to −1.8°C, despite their relative low altitude (322–810 m, Table 3.2a).
South America is characterized largely by low to medium PC2 scores, with rel- atively high rainfall variability (82.2–116.3%), and low summer temperatures (14.7–21.4°C) and altitudes (189–1222 m, Table 3.2b).
In Australia, the area for cultivating chickpea is characterized by intermedi- ate PC2 scores, increasing from east to west in the southern, Mediterranean-type region, and from south to north along the east coast. In the summer-dominant rainfall regions of the northern east coast, PC2 scores tend to increase from east to west. Altitudes and rainfall variability tend to be relatively low (50–612 m, 27.9–119.8%) within their respective PC2 categories.
This climate analysis demonstrates that while the world’s chickpea- growing regions can be grouped into four coarse rainfall and temperature categories (Mediterranean rainfall distribution, cool or warm climate; summer-dominant rainfall, cool or warm climate), there is a climatic diversity within and between regions captured by the sliding scale of PC1 and PC2 scores (Fig. 3.2a and b).
The rest of this chapter will discuss chickpea adaptation from an agroclimatic perspective.