Table 1. Risk factors (RFs) used in this study, their derivation, and the motivation for including them.

Risk  RF Name Properties Description
Introduction
  1 Immigration Dynamic Any bird arriving in the ecosystem can potentially carry a strain of influenza from another ecosystem. We quantified immigration conservatively, as the difference between the number of birds observed in a count at time t and a count in the same location at time t-1. Negative changes (emigration) were entered as zeros. No value for this RF exists for the first count session by definition.
  2 Related AIV Risk Non-dynamic Birds can introduce different pathogens from different areas. The local risk (notably for farmers) is therefore related to the type of strains that are likely to be introduced. We defined four movement patterns and ranked them according to the associated risk of introducing different strains of AIV in Chivero-Manyame ecosystem: a) resident species, associated with risk value of 0; b) species nomadic in Southern Africa, associated with a risk value of 1 (HPAI H5N1 has not been recorded in African South of the equator -OIE 2009- but other HP strains have been recorded) or 0 for H5N1 risk; c) Trans-equatorial migrants, with a risk value of 2 as HPAI H5N1 is now endemic in some African countries and outbreaks occurred in 11 countries (OIE 2009); and d) paleartic migrants, associated with a risk value of 3 because of the high number of HPAI H5N1 outbreaks and reported prevalence of LPAI is higher than in Africa(Olsen et al. 2006). For species that evidence several different strategies, as with the wood sandpiper Tringa glareola which has both migratory and resident populations (Underhill et al. 1999, Hockey et al. 2005), a mean between the two relevant coefficients was taken.
Maintenance
  3 Abundance Dynamic Total number of bird observed per species, obtained by summing numbers seen during the 60 counts. Note that since only 56 counts were done during the first count session (May 2007), we multiplied the numbers of birds recorded during this session by 60/56 for full comparability.
  4 Gregariousness Dynamic The degree of intra-species aggregation. Aggregation facilitates pathogen transmission and maintenance in the species. For each species we calculated the average group size observed across all study sites.
  5 Mixing Dynamic The degree of inter-specific aggregation, which facilitates pathogen transmission from one species to another. We estimated the degree of mixing for each species and for each count session as the ratio of the number of species observed on the same sites and at the same time, divided by the total number of species counted during the 60 counts of the count session (total species diversity measured during a count session).
  6 Percentage of juveniles in the population Dynamic Juveniles are considered to play a role in the epidemiology of AIV once they have joined the adult population (i.e., after fledging). Juveniles are also thought to remain epidemiologically naïve in the population for about 2 months (Stallknecht et al. 1990b). To capture this risk, we used Roberts’ Birds of Southern Africa (Hockey et al. 2005) to provide data on: a) clutch size; b) breeding success; and c) laying dates for the 254 species in the data set. Using a simple population model assuming constant mortality in adults (4,5% per month) and a decreasing mortality in juveniles (starting at 40% in month 1 and reaching 4,5% at 6 months), and integrating the reproductive information, the percentage of juveniles in the population was estimated by month. Incubation and fledging periods were added to determine the delay between egg laying and the entry of juveniles into the population. We considered juveniles for each species to be susceptible to AIV infection based on their naïve immunological status but despite lack of information on susceptibility for most African species.
  7 Feeding habits Non-dynamic Transmission of AIV strains in surface water is possible (Stallknecht et al. 1990a, Brown et al. 2007b), and we identified four feeding behaviors that were ranked according to the risk of birds being infected with AI during their feeding activities. They include: (0) feeding on insects on flight, seeds, nectar or fruits; (1) feeding on birds, small vertebrates, or insects close to water; (2) diving or feeding on insects gleaned from open water; and (3) dabbling, gleaning on or near surface and subsurface vegetation, or probing.