Trop Anim Health Prod (2015) 47:1005–1016 DOI 10.1007/s11250-015-0817-4
REGULAR ARTICLES
Characterization of smallholder pig breeding practices within a rural commune of North Central Vietnam Nahoko Ieda 1 & Quang Van Bui 2 & Nga Thi Duong Nguyen 2 & Lucy Lapar 3 & Karen Marshall 3
Received: 21 September 2014 / Accepted: 31 March 2015 / Published online: 7 May 2015 # Springer Science+Business Media Dordrecht 2015
Abstract This case study focused on a pig production system in a rural area of North Central Vietnam, with a focus on describing household pig breeding practices and estimating herd demographic parameters, particularly on reproduction. One hundred five households undertaking small-scale piglet production were surveyed, with information gathered on 3268 individual pigs. Pig keeping contributed variably to the overall household livelihood portfolio, with female household members as the main decision makers, contributors to labor, and beneficiaries of income from the pig enterprise. All households kept between one and four young or adult sows, with 69 % of these sows of a local breed type (predominantly Mong Cai), 28 % a cross between a local sow and an exotic sow (predominantly Large White), and the remainder (3 %) as exotic sows. Eighty-eight percent of the piglets produced were cross-bred, while 12 % were local breed. No adult males were kept by the surveyed households, reflecting the common use of artificial insemination for mating purposes. The most common breeding system practiced—the keeping of Mong Cai females and production of cross-bred piglets—capitalizes on the small body size and high fecundity of the sows and the fast growth rate and leanness of the cross-bred piglets. The survey tool used, which was based on farmer recall of events over the preceding 12-month period, appeared to give reasonable results although some recall bias could be detected. This case * Karen Marshall [email protected] 1
Graduate School of Bioagricultural Sciences, Nagoya University, Nagoya, Japan
2
Hanoi University of Agriculture, Hanoi, Vietnam
3
The International Livestock Research Institute, P.O. Box 3070900100, Nairobi, Kenya
study will serve as an entry point to planned broader scale characterization and development of pig breeding systems in North Central Vietnam. Keywords Smallhold pig production . North Central Vietnam . Mong Cai . Household survey . Animal-level survey
Introduction Pigs are one of the most important livestock species in Vietnam, which had a national herd of 27.0 million in 2011 (FAOSTAT) and where fresh pork is the favored meat of domestic consumers (ILRI et al. 2011; Lapar et al. 2012). Over recent years, and particularly during the period 1990 to 2010, there was a significant increase in national pork production, both due to an increase in the national herd (for example, from 12.3 million head in 1990 to 27.4 million head in 2005, FAOSTAT) and increases in yield (for example, from about 70 kg per head in 2001 to 100 kg per head in 2007) (ILRI et al. 2011). This growth was driven by both human population increase and an increased per capita demand for pork, particularly by a growing middle class (Roessler et al. 2009). Despite this expansion in the pig sector, as well as national livestock development strategy promoting commercialized pig farms (Lemke et al. 2007), the major suppliers of pork are smallhold farmers, typically with one to five sows, who collectively keep an estimated 80 % of the national pig population (FAO 2005; ILRI et al. 2011). A study on smallholder competiveness (Lapar et al. 2012) indicated that smallhold pig production was currently competitive and that smallholders are likely to remain the dominant pork suppliers. Smallhold pig keepers within Vietnam use a variety of breed types, including local breeds (such as the I and Mong Cai), exotics breeds (such as Large White, Landrace, Berkshire, and Duroc), and crosses between these. Local pigs
1006
presently comprise about one-quarter of the national pig herd and are mainly found in uplands and remote and rural areas (Huyen et al. 2005). Exotic breeds have been imported from America, Europe, and elsewhere, with their spread supported by the common use of artificial insemination, their promotion by various organizations including development agencies, and the decentralized structure of the Vietnamese pig breeding system (Huyen et al. 2005). The particular breed and breed crossing systems used within Vietnam vary across regions, as well as between households within a region, and in some cases are not well characterized. Similarly, there are significant knowledge gaps in the relative performance of different pig breed types within the various production systems, from both a socioeconomic and productivity perspective. This case study focused on a pig production system in a rural area of North Central Vietnam. The primary aims of the study were as follows: to characterize the breeding system(s) currently being used by smallhold pig keepers within the study site, as well as other pig management practices; to estimate selected demographic parameters (such as reproductive rates) for the different breed types kept; and to test the utility of a survey tool developed for other species (cattle, sheep, and goat; Lesnoff et al. 2010) to obtain the necessary field data for such an analysis in a rapid and low-resource manner. This study serves as an entry point to a broader scale characterization of pig breeding systems within North Central Vietnam, aimed at identifying breeding interventions beneficial to pig smallholders and other actors along the pork value chain.
Material and methods Selection of the study site and households This case study was conducted in the Van Son commune of the Do Luong District within Nghe An Province, in North Central Vietnam (Fig. 1). The Nghe An Province was of interest due to the relatively high number of livestock farms in the province (about 54 % of total farms in comparison to 29 % for the North Central and Central Coast areas; 2011 data from http://www.gso.gov.vn/), and the Van Son commune was of interest because it is a rural commune with a local reputation as a piglet production site. The Van Son commune has a human population of about 4700 and is located about 3 km from the urban center of Do Luong. Government records for 2010 to 2011 indicated that out of the 1257 households in the Van Son commune, there were 526 households undertaking small-scale piglet production (where a small-scale piglet producer was defined as owing between one and five reproductive sows). Of these 526 households, 105 households were randomly selected and surveyed. The households were located in 13 villages, with between 1 and 14 households surveyed per village (3 villages had between 1
Trop Anim Health Prod (2015) 47:1005–1016
and 5 households surveyed, 5 villages had between 6 and 10 households surveyed, and 5 villages had between 10 and 15 households surveyed). Of the 105 households, 84 % produced piglets to sell at the age of weaning (for the remainder of the report referred to as farrow to wean households), 1 % produced piglets to sell at the age of slaughter (for the remainder of the report referred to as farrow to finish households), and 15 % produced some piglets to sell at the age of weaning and others at the age of slaughter (a combination of farrow to wean and farrow to finish). Survey tools and implementation Two survey tools were utilized, a household-level survey and an animal-level survey, as described in more detail below. The surveys were implemented between 11th and 23rd of December 2012, in the local language. In all cases, the survey respondent was a household member knowledgeable on pig keeping within the household. Eighty-three percent of the respondents were female and 17 % male. Nineteen percent of the respondents were household heads and 81 % other adult household members. As in Vietnam, all livestock are considered to be owned by the family, rather than individual family members, and all respondents could be considered the pig owners. The majority (90 %) of the respondents indicated that the 12-month recall period of the animal-level survey was Bworse than an average year^ for keeping pigs, primarily due to low market price (indicated by 86 % of those considering the year to be worse than average). The household-level survey was aimed at collecting basic information on the households as well as pig management practices. Specific areas probed were as follows: household livelihood activities, household income for 2012, and the contribution of pig keeping to this income; pig feeding, healthcare, and housing practices; pig ownership, decision making, and labor responsibilities; and the reasons underpinning breed preference. The majority of questions were closed-ended (i.e., with a predetermined set of options to choose from) though some questions were open-ended (such as that on breed preference). For results given as a percentage of households, the denominator used was the number of informative households (i.e., those of the relevant category who answered that particular survey question). The animal-level survey was adapted from that presented in Lesnoff et al. (2010) which was designed to estimate various demographic parameters in tropical ruminant livestock populations (Lesnoff et al. 2010). This is the first report of this tool being utilized on a pig population. It should be noted that parameters estimated using this methodology are approximate and should be interpreted with care. This is because (a) the data is based on respondent recall and, thus, likely to contain a certain level of error which may bias the results; (b) some parameters are estimated from a limited amount of data due to the low representation of pigs of certain types within the study site;
Trop Anim Health Prod (2015) 47:1005–1016
1007
Fig. 1 Map of the Do Luong District in Vietnam. Red dot indicates the present study site and gray area indicates the studied province
Van Son Commune, Vietnam, Study Site
Nghe An Van Son
Legend Study Site Study Province Province Boundary Country Boundary
0
and (c) the results relate to the specific environmental conditions of the 12-month period surveyed, rather than being an average over many years. The best option for validation of the demographic parameters presented here, if required, is a long-term longitudinal survey. From the results presented, it appears that age of pig and litter size are subject to some recall error and that parity may often have been given as double the animals age (from the assumption that pigs will have two litters a year). Additionally, it was not possible to obtain uncensored estimates of abortion rates and stillbirth rates, due to the practice of removing females that abort or have stillbirths from the herds. Further use of this tool is thus recommended for understanding household breed use and breeding practices and in cases where estimates of these demographic parameters will suffice. In
50 100
KM 200
situations where more accurate estimation of the herd demographic parameters is required, including for more detailed breed type comparison, longitudinal monitoring is recommended. The survey respondent was interviewed in relation to the household’s pig herd in the presence of the animals. These animals were enumerated for a number of basic data (such as breed, sex, age, and parity). The respondent was also asked to recall all demographic events (such as births, deaths, sales, purchases) that occurred in the 12-month period prior to the date of the survey. A full list of the data recorded is given in Table 1. In total, information was collected on 3268 animals, as 358 animals present at the time of survey and 2910 exiting in the 12-month period prior to the survey (across the combined household herds).
1008 Table 1
Trop Anim Health Prod (2015) 47:1005–1016 Data recorded for the animal-level survey
Level
Data
For each survey
Generic: survey date, survey location (site, village) Interviewee: name, gender, type (household head, other household member), livestock owner (yes, no)
For each animal present in the herd/flock at the time of survey
Seasonal effect over the last 12 months compared to the previous 5 years (choices of Bworse than average,^ Baverage,^ Bbetter than average^) Breed type (whether pure-bred or cross-breed and based on farmer’s designation) Sex (male, female, unknown) Born in herd (yes, no) Born from AI (yes, no) Age (in months to 1 year of age, then in half-yearly intervals)
For each female present in the herd/flock at the time of survey
Number of lifetime parturitions (where a parturition is defined as the process of giving birth) Number of parturitions over the last 12 months For each parturition over the last 12 months, the number of offspring born alive For each parturition over the last 12 months, the number of offspring stillborn Number of lifetime abortions Number of abortions over the last 12 months Number of lifetime artificial insemination attempts Number of lifetime conceptions to artificial insemination
For each animal that has entered the herd over the last 12 months
For each animal that has exited the herd over the last 12 months
Breed type (as above) Sex (as above) Age (as above) Type of entry (purchase or barter, arrival in loan, returned from loan, gift, inheritance or dowry) Breed type (as above) Sex (as above) Age (as above) Type of exit (natural death, slaughter, sale or barter, departure in loan, sending back of loan, gift or dowry, theft) Type of slaughter or sale (ordinary, emergency due to disease, emergency due to lack of feed, emergency due to traumatism)
This data was used to (a) describe the state of the herds at the time of survey (such as herd size, sex by age structure), (b) characterize the breeding system(s) used, and (c) estimate various demographic parameters as listed in Table 2 as well as other demographic descriptors (for example, proportion of households selling or purchasing pigs over the 12-month recall period). Rates in Table 2 estimated as annual instantaneous hazard rates (h) were calculated as h=m/T, where m is the number of events (for example, parturition events) and T is the total time at risk, for a particular category of animals (for example, females of a particular age class) over the last 12 months. T is approximated by averaging the estimated number of animals 12 months ago (nt−1) and the current number of animals in that category (nt). nt−1 for a particular age class (i) is calculated as: nt1;i ¼ nt;iþ1 ment;i mexi;i =2 ment;iþ1 mexi;iþ1 =2 ; where ment and mexi are entries and exists into age class i in the last 12 months, respectively. nt,i is known from the data. An underlying assumption of this calculation is that there is a
uniform distribution of demographic events over time (Lesnoff et al. 2010). It should be noted that parameters estimated using this methodology are approximate and should be interpreted with care. This is because (a) the data is based on respondent recall and, thus, likely to contain a certain level of error which may bias the results; (b) some parameters are estimated from a limited amount of data due to the low representation of pigs of certain types within the study site; and (c) the results relate to the specific environmental conditions of the 12-month period surveyed, rather than being an average over many years. The best option for validation of the demographic parameters presented here, if required, is a long-term longitudinal survey.
Respondents’ perception of the 12-month recall period The majority (90 %) of the respondents indicated that the 12month recall period of the animal-level survey was Bworse
Trop Anim Health Prod (2015) 47:1005–1016 Table 2 Demographic parameters estimated from the animal-level survey data
1009
Natural rates Abortion rate
Annual instantaneous hazard rate of abortion (expected number of abortions per female when spending all the year in the herd; an abortion is a gestation that has not reached its term). Also calculated over the complete reproductive female histories (as the slope of the regression line fitted between ages and lifetime abortions of the females present in the herd)
Parturition rate
Annual instantaneous hazard rate of parturition (expected number of parturitions per female when spending all the year in the herd). Also calculated over the complete reproductive female histories (as the slope of the regression line fitted between ages and parities of the females present in the herd) Average number of offspring (stillborn or born alive) per parturition
Prolificacy rate Stillbirth rate
Probability that an offspring is a stillborn (stillbirth is not included in the mortality rate, which only concerns animals born alive)
Mortality rate
Annual instantaneous hazard rate of natural death (natural death refers to all types of death except slaughtering) Number of lifetime conceptions to artificial insemination/number of lifetime artificial insemination attempts
Artificial insemination success rate
than an average year^ for keeping pigs, primarily due to low market price (indicated by 86 % of those considering the year to be worse than average). Grouping of animals for data analysis For each animal, a breed type was assigned according to the respondent’s designation. The breed types reported are shown in Table 3, along with the total number of pigs of that breed type. Because some breed types had few animals and because of uncertainty around the designation of the breed types, the breed designation used Table 3
in the remainder of this report were simplified to local, cross-breed (between local and exotic), exotic, and unknown, as also described in Table 3. Pigs were also grouped by a combination of age class and sex. For age class, adults were defined as being 1 year of age or older, young were defined as being between 2 months and 1 year of age, and piglets were defined as being 2 months of age or younger (as piglets are generally weaned at 2 months of age). Sex of piglets was recorded as male, female, or unknown. Note that piglet sex was generally unknown, due to the common practice of first sexing piglets at weaning age.
Breed types as reported by the survey respondent, and designation used in this report
Breed type as reported by the survey respondent
Number of animals present in herds at the time of survey
Number of animals exiting herds over the 12-month period prior to survey
Breed designation used in this report
Total animals in dataset
I × Mong Cai Mong Cai Mong Cai × Large White Mong Cai × (Mong Cai × Large White) Large White × (Mong Cai × Large White) Duroc × (Large White x Mong Cai) Mong Cai × Large White (F2) Landrace/Pietran × (Yorkshire × Mong Cai) Blocally referred to as 3 blood cross^ Unknown exotic × (Landrace/Pietran × (Yorkshire × Mong Cai)) Mong Cai × unknown exotic Large White Duroc Large White × Duroc Mong Cai × Unknown Total
0 156 144 0 32 5 2 1
8 458 1800 50 378 17 34 0
Local
622
Cross-breed (local × exotic)
2534
0 12 5 1 0 0 358
12 47 0 29 12 65 2910
Exotic
47
Unknown
65 3268
1010
Trop Anim Health Prod (2015) 47:1005–1016
Results Livelihood profile of the surveyed household and contribution of pigs to household income The majority (96 %) of the pig-keeping households were engaged in a mixed crop-livestock system, while the remainder (4 %) kept livestock only. Of those practicing a mixed system, 25 % derived some of their livelihood from an off-farm activity (e.g., second business or being employed). All of the households kept more than one type of livestock. The most common combination was the keeping of pigs, cattle/buffalo, and chicken/ducks (81 % of households), followed by the keeping of pig and chicken/ducks only (14 % of households). The remainder of the households (5 %) kept a different combination of livestock such as with fish. One hundred one of the 105 respondents stated their annual household income in 2012, both in total and in relation to pig keeping, in the local currency of Vietnamese Dong (VND). The stated total household income averaged 22,014,500 VND (with a standard deviation of 13,367,491), equivalent to 1044 USD (634 USD), while the stated income from the pig enterprise averaged 9,644,000 (6,370,489), equivalent to 457 USD (302 USD) (assuming 1 USD=21,062.60 VND). The proportion of households deriving 0, 1–25, 26–50, 51–75, 76–99, and 100 % of their household income from pig keeping was 8, 12, 41, 20, 4, and 15 %, respectively, indicating that pig keeping had a varied contribution to the households income. As noted above, piglet sale price was considered to be low in 2012, and thus, in more Btypical^ years, the proportion of household income from pig keeping may be higher for some households. Pig herd structure
breed (69 % of all females kept), followed by the crossbreed (28 %), with few exotic breed (3 %), while the most common type of piglet present was a cross-breed (88 % of piglets), with the remaining piglets (12 %) of the local breed. No adult males were kept by the surveyed households, reflecting the common use of artificial insemination for mating purposes. The relatively few young males present in the combined herd were of the cross-breed type, 3 months of age, and all kept by households practicing a combination of Bfarrow to wean^ and Bfarrow to finish.^ The age structure of young and adult females, across the combined herds, is shown in Fig. 2. The majority of females (92 %) were
REGULAR ARTICLES
Characterization of smallholder pig breeding practices within a rural commune of North Central Vietnam Nahoko Ieda 1 & Quang Van Bui 2 & Nga Thi Duong Nguyen 2 & Lucy Lapar 3 & Karen Marshall 3
Received: 21 September 2014 / Accepted: 31 March 2015 / Published online: 7 May 2015 # Springer Science+Business Media Dordrecht 2015
Abstract This case study focused on a pig production system in a rural area of North Central Vietnam, with a focus on describing household pig breeding practices and estimating herd demographic parameters, particularly on reproduction. One hundred five households undertaking small-scale piglet production were surveyed, with information gathered on 3268 individual pigs. Pig keeping contributed variably to the overall household livelihood portfolio, with female household members as the main decision makers, contributors to labor, and beneficiaries of income from the pig enterprise. All households kept between one and four young or adult sows, with 69 % of these sows of a local breed type (predominantly Mong Cai), 28 % a cross between a local sow and an exotic sow (predominantly Large White), and the remainder (3 %) as exotic sows. Eighty-eight percent of the piglets produced were cross-bred, while 12 % were local breed. No adult males were kept by the surveyed households, reflecting the common use of artificial insemination for mating purposes. The most common breeding system practiced—the keeping of Mong Cai females and production of cross-bred piglets—capitalizes on the small body size and high fecundity of the sows and the fast growth rate and leanness of the cross-bred piglets. The survey tool used, which was based on farmer recall of events over the preceding 12-month period, appeared to give reasonable results although some recall bias could be detected. This case * Karen Marshall [email protected] 1
Graduate School of Bioagricultural Sciences, Nagoya University, Nagoya, Japan
2
Hanoi University of Agriculture, Hanoi, Vietnam
3
The International Livestock Research Institute, P.O. Box 3070900100, Nairobi, Kenya
study will serve as an entry point to planned broader scale characterization and development of pig breeding systems in North Central Vietnam. Keywords Smallhold pig production . North Central Vietnam . Mong Cai . Household survey . Animal-level survey
Introduction Pigs are one of the most important livestock species in Vietnam, which had a national herd of 27.0 million in 2011 (FAOSTAT) and where fresh pork is the favored meat of domestic consumers (ILRI et al. 2011; Lapar et al. 2012). Over recent years, and particularly during the period 1990 to 2010, there was a significant increase in national pork production, both due to an increase in the national herd (for example, from 12.3 million head in 1990 to 27.4 million head in 2005, FAOSTAT) and increases in yield (for example, from about 70 kg per head in 2001 to 100 kg per head in 2007) (ILRI et al. 2011). This growth was driven by both human population increase and an increased per capita demand for pork, particularly by a growing middle class (Roessler et al. 2009). Despite this expansion in the pig sector, as well as national livestock development strategy promoting commercialized pig farms (Lemke et al. 2007), the major suppliers of pork are smallhold farmers, typically with one to five sows, who collectively keep an estimated 80 % of the national pig population (FAO 2005; ILRI et al. 2011). A study on smallholder competiveness (Lapar et al. 2012) indicated that smallhold pig production was currently competitive and that smallholders are likely to remain the dominant pork suppliers. Smallhold pig keepers within Vietnam use a variety of breed types, including local breeds (such as the I and Mong Cai), exotics breeds (such as Large White, Landrace, Berkshire, and Duroc), and crosses between these. Local pigs
1006
presently comprise about one-quarter of the national pig herd and are mainly found in uplands and remote and rural areas (Huyen et al. 2005). Exotic breeds have been imported from America, Europe, and elsewhere, with their spread supported by the common use of artificial insemination, their promotion by various organizations including development agencies, and the decentralized structure of the Vietnamese pig breeding system (Huyen et al. 2005). The particular breed and breed crossing systems used within Vietnam vary across regions, as well as between households within a region, and in some cases are not well characterized. Similarly, there are significant knowledge gaps in the relative performance of different pig breed types within the various production systems, from both a socioeconomic and productivity perspective. This case study focused on a pig production system in a rural area of North Central Vietnam. The primary aims of the study were as follows: to characterize the breeding system(s) currently being used by smallhold pig keepers within the study site, as well as other pig management practices; to estimate selected demographic parameters (such as reproductive rates) for the different breed types kept; and to test the utility of a survey tool developed for other species (cattle, sheep, and goat; Lesnoff et al. 2010) to obtain the necessary field data for such an analysis in a rapid and low-resource manner. This study serves as an entry point to a broader scale characterization of pig breeding systems within North Central Vietnam, aimed at identifying breeding interventions beneficial to pig smallholders and other actors along the pork value chain.
Material and methods Selection of the study site and households This case study was conducted in the Van Son commune of the Do Luong District within Nghe An Province, in North Central Vietnam (Fig. 1). The Nghe An Province was of interest due to the relatively high number of livestock farms in the province (about 54 % of total farms in comparison to 29 % for the North Central and Central Coast areas; 2011 data from http://www.gso.gov.vn/), and the Van Son commune was of interest because it is a rural commune with a local reputation as a piglet production site. The Van Son commune has a human population of about 4700 and is located about 3 km from the urban center of Do Luong. Government records for 2010 to 2011 indicated that out of the 1257 households in the Van Son commune, there were 526 households undertaking small-scale piglet production (where a small-scale piglet producer was defined as owing between one and five reproductive sows). Of these 526 households, 105 households were randomly selected and surveyed. The households were located in 13 villages, with between 1 and 14 households surveyed per village (3 villages had between 1
Trop Anim Health Prod (2015) 47:1005–1016
and 5 households surveyed, 5 villages had between 6 and 10 households surveyed, and 5 villages had between 10 and 15 households surveyed). Of the 105 households, 84 % produced piglets to sell at the age of weaning (for the remainder of the report referred to as farrow to wean households), 1 % produced piglets to sell at the age of slaughter (for the remainder of the report referred to as farrow to finish households), and 15 % produced some piglets to sell at the age of weaning and others at the age of slaughter (a combination of farrow to wean and farrow to finish). Survey tools and implementation Two survey tools were utilized, a household-level survey and an animal-level survey, as described in more detail below. The surveys were implemented between 11th and 23rd of December 2012, in the local language. In all cases, the survey respondent was a household member knowledgeable on pig keeping within the household. Eighty-three percent of the respondents were female and 17 % male. Nineteen percent of the respondents were household heads and 81 % other adult household members. As in Vietnam, all livestock are considered to be owned by the family, rather than individual family members, and all respondents could be considered the pig owners. The majority (90 %) of the respondents indicated that the 12-month recall period of the animal-level survey was Bworse than an average year^ for keeping pigs, primarily due to low market price (indicated by 86 % of those considering the year to be worse than average). The household-level survey was aimed at collecting basic information on the households as well as pig management practices. Specific areas probed were as follows: household livelihood activities, household income for 2012, and the contribution of pig keeping to this income; pig feeding, healthcare, and housing practices; pig ownership, decision making, and labor responsibilities; and the reasons underpinning breed preference. The majority of questions were closed-ended (i.e., with a predetermined set of options to choose from) though some questions were open-ended (such as that on breed preference). For results given as a percentage of households, the denominator used was the number of informative households (i.e., those of the relevant category who answered that particular survey question). The animal-level survey was adapted from that presented in Lesnoff et al. (2010) which was designed to estimate various demographic parameters in tropical ruminant livestock populations (Lesnoff et al. 2010). This is the first report of this tool being utilized on a pig population. It should be noted that parameters estimated using this methodology are approximate and should be interpreted with care. This is because (a) the data is based on respondent recall and, thus, likely to contain a certain level of error which may bias the results; (b) some parameters are estimated from a limited amount of data due to the low representation of pigs of certain types within the study site;
Trop Anim Health Prod (2015) 47:1005–1016
1007
Fig. 1 Map of the Do Luong District in Vietnam. Red dot indicates the present study site and gray area indicates the studied province
Van Son Commune, Vietnam, Study Site
Nghe An Van Son
Legend Study Site Study Province Province Boundary Country Boundary
0
and (c) the results relate to the specific environmental conditions of the 12-month period surveyed, rather than being an average over many years. The best option for validation of the demographic parameters presented here, if required, is a long-term longitudinal survey. From the results presented, it appears that age of pig and litter size are subject to some recall error and that parity may often have been given as double the animals age (from the assumption that pigs will have two litters a year). Additionally, it was not possible to obtain uncensored estimates of abortion rates and stillbirth rates, due to the practice of removing females that abort or have stillbirths from the herds. Further use of this tool is thus recommended for understanding household breed use and breeding practices and in cases where estimates of these demographic parameters will suffice. In
50 100
KM 200
situations where more accurate estimation of the herd demographic parameters is required, including for more detailed breed type comparison, longitudinal monitoring is recommended. The survey respondent was interviewed in relation to the household’s pig herd in the presence of the animals. These animals were enumerated for a number of basic data (such as breed, sex, age, and parity). The respondent was also asked to recall all demographic events (such as births, deaths, sales, purchases) that occurred in the 12-month period prior to the date of the survey. A full list of the data recorded is given in Table 1. In total, information was collected on 3268 animals, as 358 animals present at the time of survey and 2910 exiting in the 12-month period prior to the survey (across the combined household herds).
1008 Table 1
Trop Anim Health Prod (2015) 47:1005–1016 Data recorded for the animal-level survey
Level
Data
For each survey
Generic: survey date, survey location (site, village) Interviewee: name, gender, type (household head, other household member), livestock owner (yes, no)
For each animal present in the herd/flock at the time of survey
Seasonal effect over the last 12 months compared to the previous 5 years (choices of Bworse than average,^ Baverage,^ Bbetter than average^) Breed type (whether pure-bred or cross-breed and based on farmer’s designation) Sex (male, female, unknown) Born in herd (yes, no) Born from AI (yes, no) Age (in months to 1 year of age, then in half-yearly intervals)
For each female present in the herd/flock at the time of survey
Number of lifetime parturitions (where a parturition is defined as the process of giving birth) Number of parturitions over the last 12 months For each parturition over the last 12 months, the number of offspring born alive For each parturition over the last 12 months, the number of offspring stillborn Number of lifetime abortions Number of abortions over the last 12 months Number of lifetime artificial insemination attempts Number of lifetime conceptions to artificial insemination
For each animal that has entered the herd over the last 12 months
For each animal that has exited the herd over the last 12 months
Breed type (as above) Sex (as above) Age (as above) Type of entry (purchase or barter, arrival in loan, returned from loan, gift, inheritance or dowry) Breed type (as above) Sex (as above) Age (as above) Type of exit (natural death, slaughter, sale or barter, departure in loan, sending back of loan, gift or dowry, theft) Type of slaughter or sale (ordinary, emergency due to disease, emergency due to lack of feed, emergency due to traumatism)
This data was used to (a) describe the state of the herds at the time of survey (such as herd size, sex by age structure), (b) characterize the breeding system(s) used, and (c) estimate various demographic parameters as listed in Table 2 as well as other demographic descriptors (for example, proportion of households selling or purchasing pigs over the 12-month recall period). Rates in Table 2 estimated as annual instantaneous hazard rates (h) were calculated as h=m/T, where m is the number of events (for example, parturition events) and T is the total time at risk, for a particular category of animals (for example, females of a particular age class) over the last 12 months. T is approximated by averaging the estimated number of animals 12 months ago (nt−1) and the current number of animals in that category (nt). nt−1 for a particular age class (i) is calculated as: nt1;i ¼ nt;iþ1 ment;i mexi;i =2 ment;iþ1 mexi;iþ1 =2 ; where ment and mexi are entries and exists into age class i in the last 12 months, respectively. nt,i is known from the data. An underlying assumption of this calculation is that there is a
uniform distribution of demographic events over time (Lesnoff et al. 2010). It should be noted that parameters estimated using this methodology are approximate and should be interpreted with care. This is because (a) the data is based on respondent recall and, thus, likely to contain a certain level of error which may bias the results; (b) some parameters are estimated from a limited amount of data due to the low representation of pigs of certain types within the study site; and (c) the results relate to the specific environmental conditions of the 12-month period surveyed, rather than being an average over many years. The best option for validation of the demographic parameters presented here, if required, is a long-term longitudinal survey.
Respondents’ perception of the 12-month recall period The majority (90 %) of the respondents indicated that the 12month recall period of the animal-level survey was Bworse
Trop Anim Health Prod (2015) 47:1005–1016 Table 2 Demographic parameters estimated from the animal-level survey data
1009
Natural rates Abortion rate
Annual instantaneous hazard rate of abortion (expected number of abortions per female when spending all the year in the herd; an abortion is a gestation that has not reached its term). Also calculated over the complete reproductive female histories (as the slope of the regression line fitted between ages and lifetime abortions of the females present in the herd)
Parturition rate
Annual instantaneous hazard rate of parturition (expected number of parturitions per female when spending all the year in the herd). Also calculated over the complete reproductive female histories (as the slope of the regression line fitted between ages and parities of the females present in the herd) Average number of offspring (stillborn or born alive) per parturition
Prolificacy rate Stillbirth rate
Probability that an offspring is a stillborn (stillbirth is not included in the mortality rate, which only concerns animals born alive)
Mortality rate
Annual instantaneous hazard rate of natural death (natural death refers to all types of death except slaughtering) Number of lifetime conceptions to artificial insemination/number of lifetime artificial insemination attempts
Artificial insemination success rate
than an average year^ for keeping pigs, primarily due to low market price (indicated by 86 % of those considering the year to be worse than average). Grouping of animals for data analysis For each animal, a breed type was assigned according to the respondent’s designation. The breed types reported are shown in Table 3, along with the total number of pigs of that breed type. Because some breed types had few animals and because of uncertainty around the designation of the breed types, the breed designation used Table 3
in the remainder of this report were simplified to local, cross-breed (between local and exotic), exotic, and unknown, as also described in Table 3. Pigs were also grouped by a combination of age class and sex. For age class, adults were defined as being 1 year of age or older, young were defined as being between 2 months and 1 year of age, and piglets were defined as being 2 months of age or younger (as piglets are generally weaned at 2 months of age). Sex of piglets was recorded as male, female, or unknown. Note that piglet sex was generally unknown, due to the common practice of first sexing piglets at weaning age.
Breed types as reported by the survey respondent, and designation used in this report
Breed type as reported by the survey respondent
Number of animals present in herds at the time of survey
Number of animals exiting herds over the 12-month period prior to survey
Breed designation used in this report
Total animals in dataset
I × Mong Cai Mong Cai Mong Cai × Large White Mong Cai × (Mong Cai × Large White) Large White × (Mong Cai × Large White) Duroc × (Large White x Mong Cai) Mong Cai × Large White (F2) Landrace/Pietran × (Yorkshire × Mong Cai) Blocally referred to as 3 blood cross^ Unknown exotic × (Landrace/Pietran × (Yorkshire × Mong Cai)) Mong Cai × unknown exotic Large White Duroc Large White × Duroc Mong Cai × Unknown Total
0 156 144 0 32 5 2 1
8 458 1800 50 378 17 34 0
Local
622
Cross-breed (local × exotic)
2534
0 12 5 1 0 0 358
12 47 0 29 12 65 2910
Exotic
47
Unknown
65 3268
1010
Trop Anim Health Prod (2015) 47:1005–1016
Results Livelihood profile of the surveyed household and contribution of pigs to household income The majority (96 %) of the pig-keeping households were engaged in a mixed crop-livestock system, while the remainder (4 %) kept livestock only. Of those practicing a mixed system, 25 % derived some of their livelihood from an off-farm activity (e.g., second business or being employed). All of the households kept more than one type of livestock. The most common combination was the keeping of pigs, cattle/buffalo, and chicken/ducks (81 % of households), followed by the keeping of pig and chicken/ducks only (14 % of households). The remainder of the households (5 %) kept a different combination of livestock such as with fish. One hundred one of the 105 respondents stated their annual household income in 2012, both in total and in relation to pig keeping, in the local currency of Vietnamese Dong (VND). The stated total household income averaged 22,014,500 VND (with a standard deviation of 13,367,491), equivalent to 1044 USD (634 USD), while the stated income from the pig enterprise averaged 9,644,000 (6,370,489), equivalent to 457 USD (302 USD) (assuming 1 USD=21,062.60 VND). The proportion of households deriving 0, 1–25, 26–50, 51–75, 76–99, and 100 % of their household income from pig keeping was 8, 12, 41, 20, 4, and 15 %, respectively, indicating that pig keeping had a varied contribution to the households income. As noted above, piglet sale price was considered to be low in 2012, and thus, in more Btypical^ years, the proportion of household income from pig keeping may be higher for some households. Pig herd structure
breed (69 % of all females kept), followed by the crossbreed (28 %), with few exotic breed (3 %), while the most common type of piglet present was a cross-breed (88 % of piglets), with the remaining piglets (12 %) of the local breed. No adult males were kept by the surveyed households, reflecting the common use of artificial insemination for mating purposes. The relatively few young males present in the combined herd were of the cross-breed type, 3 months of age, and all kept by households practicing a combination of Bfarrow to wean^ and Bfarrow to finish.^ The age structure of young and adult females, across the combined herds, is shown in Fig. 2. The majority of females (92 %) were
