An analysis of village garden management in the Papua New Guinea highlands

A survey of approximately 100 village gardeners in the Papua New Guinea (PNG) highlands was conducted in 2005 as a preliminary step in setting up an Australian Centre for International Agricultural Research project to improve the nutrition of sweetpotato in the highlands. Sweetpotato is the main staple crop in the PNG highlands, which had a population of around 2 million (or 40% of the country’s population) at the 2000 census, with around 2–3% annual growth. Despite these high growth rates, the area under agricultural production has remained fairly static, resulting in increasing pressure on land resources. Farmers are concerned about yield decline, as sweetpotato yields from fallowed land, or land recently brought back into production as part of the recycling of gardens that occurs in the highlands, tend to be much higher (2–8 t/ha tubers) than yields from old gardens. Despite high apparent nutrient stocks and favourable carbon:nitrogen ratios in the soil, most sweetpotato tissue samples have shown low nutrient levels, in particular of nitrogen, phosphorus, potassium and boron. The survey was conducted to assemble background information on farm locations, farm practices, crop yields, and soil and plant analysis. The conceptual framework proposed to guide the analysis was that background factors (such as location, garden type, land availability, age and sex of farmer, number of children in household and sources of other income) affect management practices. These management practices (soil preparation, planting system, fertility management including fallowing, use of animals such as goats and pigs in the farming system, and crop rotations) influence outcomes such as plant and soil analyses; sweetpotato yields; plant symptoms; tuber characteristics such as weight, size, colour and cracking; presence of nematodes; and ability to grow other crops. A preliminary analysis of the survey data has been conducted and results are reported in this paper.