Appendix
Figure S1. (previous page) (A) Distribution of HPAI H5N1 virus outbreaks in Bangladesh, India, Myanmar, Thailand, Cambodia, Laos and Vietnam (data for Thailand and Vietnam differentiate records from the second wave in blue and from records that took place afterward in red) (B) Predictions from the Southeast Asia model published in Gilbert et al. (2005), with areas of good fit (blue boxes) and overestimation of risk (red boxes); (C), (D) & (E) Distribution of chicken, duck and human density (heads / km2), respectively; (F) Elevation above sea level (m); (G) Average number of crop cycles per year. The figure highlights that the previously published model (the Southeast Asia model) correctly predicts HPAI H5N1 virus risk in Bangladesh, but overestimates the risk in India, and that this is primarily caused by high cropping intensities and low elevation found in north India (long red rectangle in panel B).
Figure S2. Density of HPAI H5N1 outbreaks found in areas defined by intervals of rice cropping intensity (green), duck (red), chicken (orange), and human (blue) density (heads / km2 in log10(x+1) scale). The figures highlight similar profiles in Thailand and Vietnam with increments in cropping intensity, duck density and human population density translating into higher outbreak densities. In contrast, in India and Bangladesh, a plateau is observed in outbreak density for increasing duck density class intervals.
Figure S3. Temporal distribution of the number of GPS position per day for the wild waterfowls trajectories displayed in Fig. 2c. The blue rectangles highlight periods of stays in Bangladesh and Qinghai regions (defined by 1 & 2 in Fig. 2c). The red arrows are indicative of the timing of the nearest HPAI H5N1 outbreaks in the respective regions.
Table S1. Sources of poultry data used in this study. For each country, data on ducks and chickens were obtained at the finest available administrative unit level.
Bangladesh / Bangladesh Department of Livestock Services, Ministry of Fisheries & Livestock, Dhaka, Bangladesh, Agricultural Sample Survey of Bangladesh – 2005.
Cambodia / Department of Animal Health 2003 Census*
India / 2003 Agricultural Census, Agricultural Census Division, Ministry of Agriculture
Laos / Ministry of Agriculture and Forestry, Department of Planning, Statistics Division, Laos*
Myanmar / Myanmar Animal Census*
Thailand / Department of Livestock Development, Bangkok Thailand (X-Ray Survey in 2004)
Vietnam / 2001 Rural Agriculture and Fisheries Census 2001, Ministry of Agriculture, Hanoi, Vietnam
* Obtained through the Global Livestock and Production Atlas (GLIPHA, FAO)
Table S2. Autologistic multiple regression results for the South and Southeast Asia model and the South Asia model Results are provided from 250 bootstrapped regressions with 9 times more negatives than n positives, an autoregressive terms estimated up to a distance of 0.5 decimal degrees. Significance of individual terms is estimated from the likelihood ratio Chi-squared test.
Coef. Chi p Sig.
South & Southeast Asia Model
Intercept -6.642 ± 0.351 - -
ChDnLg 0.670 ± 0.170 17.53 <0.001 ***
DuDnLg 1.374 ± 0.187 46.20 <0.001 ***
HpopLg 0.182 ± 0.173 2.55 0.1106 n.s.
NCropC -0.557 ± 0.212 5.50 0.0190 *
Art05 59.426 ± 3.268 506.73 <0.001 ***
HasDu 2.046 ± 0.532 19.58 <0.001 ***
ChDnLg:HasDu -0.218 ± 0.210 2.31 0.1286 n.s.
DuDnLg:HasDu -1.138 ± 0.210 23.78 <0.001 ***
HpopLg:HasDu 0.238 ± 0.216 2.58 0.1083 n.s.
NCropC:HasDu 0.523 ± 0.255 4.35 0.0370 *
Art05:HasDu -34.111 ± 3.548 122.28 <0.001 ***
South Asia model
Intercept -10.189 ± 0.873 - -
ChDnLg 0.156 ± 0.223 1.01 0.3158 n.s.
DuDnLg -0.516 ± 0.469 2.01 0.1567 n.s.
HpopLg 1.449 ± 0.341 18.90 <0.001 ***
NCropC -0.877 ± 0.426 5.52 0.0188 *
Art05 63.544 ± 7.904 88.47 <0.001 ***
HasDu -0.167 ± 1.241 0.93 0.3359 n.s.
ChDnLg:HasDu 0.726 ± 0.282 4.89 0.0270 *
DuDnLg:HasDu 0.647 ± 0.498 2.22 0.1364 n.s.
HpopLg:HasDu 0.388 ± 0.441 1.46 0.2273 n.s.
NCropC:HasDu 0.931 ± 0.508 5.19 0.0228 *
Art05:HasDu -53.429 ± 9.556 46.12 <0.001 ***
Table S3. HPAI H5N1 wild bird outbreaks in the Central Asian Flyway, 2005-2010. Countries are presented from south-to-north with locations, reported dates of detection, and host species and number observed*,#
2005 / China / Qinghai Lake, Gangcha county, Qinghai Province / 15 April / Bar-headed goose (3282), great cormorant (1302), great black-headed gull (929), brown-headed gull (570), ruddy shelduck (145)
Mongolia / Erhel Lake, Alag-Ederne District & Khunt Lake, Saikhan, Bulgan / 2 August / Bar-headed goose (89), whooper swan, mute swan, Eurasian wigeon, ruddy shelduck
Russia / Chany Lake, Novosibirsk / 18 July / Great crested grebe, gadwall, mallard, ruddy shelduck, mute swan, coot, common pochard
China / Nagqu Prefecture, Tibetan Autonomous Region / 19 April / Bar-headed goose (399), ruddy shelduck
2006 / Mongolia / Saikhan soum, Bulgan Province / 4 May / Whooper swan (1)
Lake Hunt, Bulgan, / 5 June / Migratory gull, swan and geese (12)
Russia / Ubsu-Nur Lake, Ovyursky District / 15 June / ‘Wild geese’ (1622)
China / Genggahu Lake, Hainan Prefecture / 8 May / Great crested grebe (121)
2009 / Mongolia / Doitiin tsagaan Lake,
Ugii-nuur Soum / 22 May / Whooper swan (9)
Doroo nuur, Tsetserleg, Arkhangai / 1 August / Bar-headed goose (34), Carrion crow (67), ruddy shelduck (28), crane (1), common goldeneye (11), grey heron (2), herring gull (16), black kite (3), whooper swan (6), unspecified (3)
Russia / Ubsu-Nur, Ovursky District / 11 June / ‘Wild birds’ (58)
2010 / Mongolia / Ganga lake, Dariganga soum, Sukhbaatar / 3 May / Greylag geese and whooper swans (26)
* World Organization for Animal Health (OIE), Update on highly pathogenic avian influenza in animals, http://www.oie.int/downld/AVIAN%20INFLUENZA/A2009_AI.php.
# FAO (2009). Emergency Prevention System; Food and Agricultural Organization. Global animal disease information system. Food and Agriculture Organisation of the United Nations, Animal Health Service, Animal Production and Health Division. Rome, http://empres-i.fao.org/empres-i/home?p=a&l=en_US
4