Australian Agency for International Development

B. J. Allen

R. M. Bourke

OCTOBER 1997

Port Moresby

Frost and drought. Much of Papua New Guinea is currently affected by a major drought which is one of the most prolonged and severe over the past 100 years. This has been caused by temporary changes associated with the El Nino climatic pertubation. The drought is at least as severe as the major droughts of the late-1890s, mid-1910s and early-1940s. It commenced as early as January 1997 in some locations and became widespread in about July. In limited parts of PNG, the rainfall has been below normal, but these places are not affected by drought. In some places, such as in South Bougainville, above average rainfall has been received. There are indications that the drought will continue for at least some months, but this cannot be predicted with any accuracy. Frost has been recorded in many highland locations and severe frosts have caused widespread devastation at high altitude locations (above 2200 m) in parts of Enga, Southern Highlands, Western Highlands and Central Provinces. Frosts are continuing.

Assessment. An objective assessment of the impact of the drought and frosts was conducted by 13 teams. The authors led one team each and the others were led by Papua New Guineans and PNG based expatriates, mostly agriculturalists from DAL. The teams assessed the impact of drought and frost on village food and water supply in all 19 provinces and most districts. The teams were composed of agriculturalists, health specialists and provincial based staff and some contained water supply specialists. It was conducted over an 18 day period in September-October and some 650 individual reports were filed. Most of the cost of this large operation was borne by AusAID. The information has been entered into a computer database set up in the Department of Provincial Affairs and Local Government in Port Moresby. A partial analysis of the information has now been completed.

Food. The assessment finds that around 80,000 people are presently in a critical life threatening situation, with no food other than ‘famine’ foods or bush foods available, and a further 70,000 are approaching this situation and will have reached it by the end of October. A further 175,000 people are eating only very small amounts of rapidly diminishing food from their gardens and are supplementing this with ‘famine’ foods and bush food. Within a month, many of them will be in the same situation as the 150,000 who effectively have nothing to eat. Another 220,000 still have food in gardens, but this food will be consumed within two months. The food that is available is probably less than half of that normally produced and is of poor quality.

Thus about 540,000 rural Papua New Guineans now have insufficient subsistence food supplies or the cash income with which to purchase food. These people need some food aid immediately. They are living in all mainland provinces and some Islands Region provinces. The worst affected provinces are in the Highlands, especially Southern Highlands, Enga and Eastern Highlands, but parts of the other mainland provinces are also badly affected. The least affected provinces are in the Islands Region. These numbers will continue to grow with each passing month. Even if widespread heavy rain falls soon, the numbers will grow because of the delay between planting and harvesting food gardens. The scale of the problems that require urgent attention are several times greater than those which occurred followed widespread frost and drought in 1972-73.

Villagers are spending their cash reserves, selling assets and changing spending patterns to buy imported food. Rice sales in 1997 are running at 25 percent above the 1996 rate, even before the delivery of significant food aid has commenced. Because the main coffee season is now finished and assets are being sold, many people will have less cash income in the coming months, particularly in the highland provinces. Cash crops including cocoa, fresh food and betel nut are being affected by the drought. It is anticipated that production of most cash crops, including coffee, will be reduced in the coming year. In many locations, villagers report that pigs and dogs are dying because of insufficient food.

Water. The majority of rural villagers in PNG are currently obtaining their drinking water from sources other than their usual ones. Many are obtaining drinking water from larger streams and rivers, some of which are heavily contaminated with human and animal waste.

The burden of carrying water is particularly severe on women and girls, who are often stressed by inadequate diet and anxiety about the health of their children. It is estimated that about 130,000 rural villagers are currently experiencing critical problems with drinking water. These problems include an almost complete lack of drinking water or access to only brackish water or access to only contaminated water. There is insufficient water to process sago in some locations, particularly in parts of Western and Gulf Provinces, the staple food there. Only a limited assessment was made of urban water supply. For some small urban locations, the situation in now critical and some government stations are closed because of the lack of water.

Fires. Fires are widespread, with smoke haze so severe at times that aircraft movements are being restricted, even for major airports. Extensive areas of grassland, forest, economic trees, sago palm and nipa stands have been burnt. In mid-October, fires are continuing in many locations. Field teams recorded many reports of damage to food gardens, garden houses, dwellings as well as to some tree crops.

Health. In many parts of PNG, villagers and local health workers report an increase in diarrhoa and skin infections. There are also many reports of increases in other diseases, including pneumonia and malaria, although it is not possible to assess these in the absence of reliable numerical information. In some places, nursing mothers report a decline in the quality or quantity of breast milk. In certain locations, villagers are attributing deaths to drought related causes, but it is not possible to adequately assess these claims.

Institutions. Throughout PNG, it is estimated that more than half of all community or primary schools are closed, either completely or they operate during the mornings only. This is because of lack of stamina by pupils and the need to search for food and water. Other institutions in the most severely affected locations are closing or are threatened with closure, mainly because of lack of water. These include hospitals, health centres, sub-health centres, aid posts, high schools and corrective institutions.

Power. No assessment was made by us of the security of power supply, which is mainly generated by four hydro-electric schemes in PNG. However, it is noted that the water held in the storage reservoirs for the two main schemes (Yonki and Sirinumu storage reservoirs) has been in steady decline since April this year, with Sirinumu now less than half full.

Recommendations. A series of recommendations are made to address the problems detailed above. These include:

• · provision of immediate food aid in certain locations;
• · facilitation of access to cash income by villagers;
• · provision of irrigation for crop production and maintenance of planting material;
• · facilitation of access to drinking water for villagers and institutions; addressing urgent health problems in the most severely affected locations;
• · restoration of infrastructure for aid delivery; and transmission of key extension messages.

‘El Nino’ has become the everyday name for an number of complex changes which occur in oceanic and atmospheric circulation across the Pacific. Water in the eastern Pacific becomes warmer than water in the western, or PNG, side of the Pacific, which is the reverse of the ‘normal’ situation. Air moving over a warm oceans picks up water vapour. This warm moist air rises, or is forced up over mountain ranges, where it cools and produces rain. Air moving over a relatively cold ocean picks up less water vapour. In a strong El Nino event, such is occurring in 1997, ocean temperatures are much lower than normal. Rather than warm moist air rising over Papua New Guinea, cool dry air is descending. The outcome is significant decline in cloud cover and rainfall. The reason that these changes or oscillations occur and why some are larger than others is not known.

These changes are reflected in changes in air pressure across the Pacific. Rather than the ‘normal’ situation, in which air pressure in Darwin is lower than air pressure in Tahiti, it becomes higher. A simple index of these pressure differences between Darwin and Tahiti is known as the Southern Oscillation Index (SOI). When the pressure in Darwin is higher than in Tahiti, the SOI is said to be in a ‘negative’ state.

Late in 1996 water in the eastern Pacific began to warm and the SOI fell unusually rapidly from around +12 in December 1996 to -25 in March 1997. In June 1997 it remained below -20 and in October it is -13. Both the rapidity of the oscillation and its magnitude suggest that this western Pacific ‘cold’ period may last into 1998. However the present El Nino models are not capable of predicting with high levels of confidence, what will happen over the next 6 to 12 months. It is known that often, at the end of an El Nino event, heavier than average rainfall occurs, as the western Pacific reverts to a warm phase. These very wet periods often create additional food supply problems in sweet potato based agricultural systems in the PNG highlands.

In Papua New Guinea the effects of these changes in Pacific circulation systems were felt almost immediately. First, rain across almost the whole country was significantly reduced from ‘normal’ falls. For example, in one of the wettest places in the country, Tabubil, only 68 mm of rain fell in July, compared to an average over 20 years of recording of 614 mm for that month. In August, only 32 mm was received compared to the average of 870 mm. No rain at all fell in the first two weeks of September. Anecdotal accounts from many part of the country suggest rainfall all but ceased in May-June, but there has been considerable variation in the actual pattern.

The outcome has been a severe reduction in crop yields, possibly by as much as 80 per cent in many places. Local springs and small streams which provided drinking water to hundreds of small communities have dried up, forcing them to travel further to larger streams for drinking water. In addition, widespread fires have occurred in grasslands and forests, from the coast to the tops of the highest mountain ranges. Fires have destroyed many existing gardens, stands of sago and in some cases houses and property.

Second, at altitudes above 2200 m above sea level (asl) cloudless night skies allow heat to radiate from the ground surface such that ground level temperatures may fall to below zero, the temperature at which water freezes. At Tambul in the Western Highlands, temperatures fell to below freezing in June and July. In the second week of August temperatures went below zero four nights in a row and in September for eight nights in succession. The coldest night in September was -2.3 o C. There have been a further three frosts in mid-October. The Tambul record will have been reflected in all the high altitude mountain basins. The recurrence, duration and coldness of these temperatures killed all top growth in sweet potato in the fields and destroyed all attempts to replant and re-establish gardens. The very low temperatures on some nights appears to have damaged tubers in the ground and rendered them unfit for consumption. Natural vegetation, including cane grasses, pandanus groves and even forest trees were killed by these frosts. In many places the drying dead vegetation has later burned.

An important feature of El Nino is that it occurs regularly every 8 to 13 years, but that severe events such as 1997 possibly occur only once in a century. Hence many people do not experience an event of such severity, and journalist publish sensationalist reports on the "worst drought in living memory". However, longer records, such as growth rates of coral on the Great Barrier Reef near the mouth of the Fly River, which can act as a surrogate for the SOI, indicate discharge from the Fly has been significantly reduced many times over thousands of years. Humans have occupied the island of New Guinea for at least 40,000 years, so human communities must have survived such events in the past. But while communities survived, many individuals did not. Accounts by observers of an event which occurred in 1941 and 1942 and oral histories of an earlier and possibly more serious event in the mid-1910s, tell of many deaths occurring at these times. Early colonial records from the 1890s and 1900s describe shipping being unable to navigate between New Guinea and New Britain because of smoke from forest fires, and of large numbers of deaths from starvation in inland areas which were then unexplored by outsiders.

It seems as if famine was not uncommon feature of pre-colonial Papua New Guinea. The first historical report of food shortage comes from Woodlark Island in about 1850. The colonial administrators have responded in the past by providing relief. The first record of aid being given is in about 1916 in New Ireland and again in 1931 at Rigo, where the cost was later taken from head taxes. In 1972-73 the Australian administration provided food up to 130,000 people in the highlands for about eight months at a cost of A$3 million (for food only) in 1972 values.

Thus the 1997 event is neither a unique one-off event, and it may not be the most severe event of this kind that has ever taken place. What is different about this event is that it is occurring in the nation state of Papua New Guinea for the first time, and the numbers of people involved are very much greater. Population growth and the widespread drought make the numbers of people in difficulties almost four times larger than in 1972.

The needs of a large scale relief operation will place great demands on the resources and structural capacity of Papua New Guinea. In normal times, rural people largely look after themselves. Now large numbers of whole communities are threatened and the ability of the state to respond effectively is critical to their survival.

It is not possible to understand the dimensions of the present crisis in PNG without some basic knowledge of how food supply systems function. The Australian National University in collaboration with DAL, has recently completed a seven year research project, funded partly by AusAID, which identifies, describes and maps agricultural systems over the whole of Papua New Guinea. Detailed descriptions of agriculture for all part of PNG are available in Provincial Working Papers, or from the Research Division of DAL in Port Moresby. Information is held in a computer database, linked to a database containing detailed information on natural resources. The mapping scale is 1:500 000.

Approximately 65 per cent of people in PNG depend upon sweet potato as a staple food and a further 15 per cent depend upon sago as a staple. Sago is not grown in gardens, but is harvested from planted and wild stands. Palms take 12-15 years to mature. Large amounts of water are required to process starch from the trunk of the palm. The next most important crop is banana. Banana stands are increasingly cultivated for between 5 and ten years, before land is left in fallow.

The large highlands populations depend upon the continuous cultivation of sweet potato. Planting and harvesting occurs continuously. Thus any disruptions to planting, or to plants at various times from planting, results in disruptions to food supply some months later and the time taken for food supply to recover from a disruption is proportional to the severity of the disruption and the maturation time of sweet potato. At altitudes up 1500 m, sweet potato is ready for harvesting in about 5-6 months. Above that altitude, time to harvest becomes progressively longer until above 2700 m it is between 9-12 months.

An important feature of the highlands sweet potato systems is the number of pigs which are supported. For the entire region, the human to pig ratio is about 1:1. About half of all sweet production is fed to pigs, usually the smallest tubers. A good indication that food supply is reduced is people eating tubers that would normally be given to pigs.

Sweet potato was probably introduced into PNG around 300 years ago. It is the plant which enable people to colonise the high altitude areas, which have been occupied only for the last 300 years. The effect of a single frost on sweet potato plants depends on the time from planting. If a frost occurs one to two months from planting, the plant becomes dormant for up to a month and then begins growing again. If frosted 5 to 6 months from planting, the dormancy is longer and the plant may die and the tubers in the ground will immediately rot. In plants frosted 9-12 months from planting, the top growth may die, but tubers in the ground may remain edible for up to two months.

A sweet potato food production system can therefore handle single frosts with few problems. However a series of frosts, particularly frosts spaced a number of weeks apart which repeatedly kills top growth and damages tubers in the ground, effectively stops all food production for up to 12 months. A series of frosts also destroys all green planting material. Following the frosts, tubers remaining in the ground can be eaten, to the point where either the supply is finished or insect damage and rotting makes them inedible. Thus for some time after a series of frosts, there is food available but inevitably a period of up to 15 months of no food is approaching. The pre-colonial response to this situation was migration to lower areas, and the gradual re-establishment of gardens with new planting materials carried up from lower areas.

Sweet potato is a hardy and drought resistant plant. Lack of rain for a time does not kill sweet potato vines, but tuber yields become increasingly depressed to the point of there being no production at all. Vine growth ceases and the plant becomes almost dormant. When rain is received, there is a period of up to 4 months during which vines are not suitable for planting. Once again, until all edible tubers in the ground are consumed, food remains available, but an inevitable period of no food production is approaching, the length of which is determined by the time it takes the vines to grow to a viable length plus the time from planting it takes the plant to produce an edible tuber.

Shifting cultivation systems involve the annual clearing of land from fallow. Fallow land may be covered in tall grasses, scrub or tall forest, depending on the length of the fallow period and environmental conditions in the area. Fallow vegetation is typically cut down, allowed to dry and burned. Planting material from last year’s garden is carried to the new garden and planted. Most families will have two gardens in operation at any time, and many will have more than two. Gardens are commonly cleared towards the end of the drier season, July to September in many parts of PNG (but importantly not all parts - in some places the seasons are reversed) and are planted to receive rain in November and December.

The drought in 1997 has significantly decreased production from gardens planted in 1996, to the extent that it appears many families have been forced to consume at least part of their planting stock, or their planting stock is significantly reduced. Many of the fires that have swept the countryside are seemingly fires lit to clear land for planting in 1997 which have escaped into dry vegetation surrounding the new garden sites. In some lowland locations, people will begin to plant their 1997-98 gardens now, in hope of rain in November and December. Until the first production from those gardens is available in about February or March 1998, assuming rain is received, they will have very little or no food available from their drought affected 1996-1997 gardens.

A wide range of trees are grown in PNG for their leaves, fruits and nuts. In the past, the nuts in particular have been critical seasonal components in the food supply system especially in places with strong dry seasons. These include in the highlands three species of pandanus which are very important. In the lowlands coconut is the most important and obvious; other include Polynesian chestnut, breadfruit, sea almond, galip nut (Canarium indicum), okari nut (Terminalia kaernbachii), Pometia pinnata and Castanopsis accuminatissima. Little is known about productive performance of these trees under severe drought conditions, but in previous food shortages, they have proved critical in keeping people alive until garden foods have become available again.

Many PNG communities have foods that are not normally eaten, but that are known can be eaten during times of famine. They include Pueraria lobata, varieties of "wild" yams (Dioscorea spp), non-planted taro-like species, green pawpaw fruit, leaves and fruit of various figs (Ficus spp), stems of Chinese taro (Xanthosoma), various green leaves, ferns and other unidentified plants. People also commonly talk about eating the basal portion of banana stems. During this assessment, these plants and others were recorded as being eaten by some villagers.

Many of these foods are unpalatable and cause mild to severe indigestion and diarrhea. Where people are eating these foods, all other sources of food are likely to be very much reduced or non-existent.

Under normal circumstances, most rural communities in PNG collect drinking water from numerous very localised sources. Very few communities use larger streams or rivers as sources of drinking water. People are acutely aware of the propensity of larger streams and rivers to be contaminated by human and animal wastes, and importantly by pollution from supernatural sources. In some areas, drinking water is restricted and of poor quality even in normal times. This is so on small islands and atolls and in limestone areas where there may be little surface water. Some coastal communities draw water from freshwater outflows which appear on fringing reefs at low tide. This water is often contaminated and brackish.

Introduced water supply schemes cannot, in general, be said to have been successful, largely for social and cultural reasons. Either no individual in the community is prepared to take responsibility for the drinking water supply for the whole community, or arguments occur over the ownership of the water. If these disputes are not resolved, they may result in the supply being sabotaged. Technical problems also plague many schemes, in particular pump breakage, problems with intakes and lack of capacity.

A drought of the dimensions of 1997 has forced people to fall back on larger streams and rivers and to carry water for considerable distances. There is little doubt many of the present sources of drinking water are contaminated (some have been shown to be by the assessment teams), but how this relates to outbreaks of diarrhea is not known. Rural people are challenged by numerous water based infectious diseases when young, and those who survive their childhood may have developed immunological resistance to the most common pathogens. Poor personal hygiene (hand washing and food preparation) which is exacerbated by a lack of washing water, is possibly a more common source of infection than contaminated water.

The 1997 drought has also reduced those water sources which were already marginal before the drought to extremely low quantity and quality source, which are strongly brackish and heavily contaminated.

A rapid assessment was undertaken to examine the impact of the drought and frosts on food supplies and water over the whole country. Drs Allen and Bourke arrived in Papua New Guinea on 21 September 1997. Together with Mr E. Sitapai, and Mr B. Wayi of the Department of Agriculture and Livestock, 13 assessment teams were organised from DAL food crops research staff at Bubia, Keravat, Aiyura and Laloki. Other team members were seconded from Ok Tedi Mining Ltd, CCRI and CCEA. All teams were accompanied by provincial based staff. The teams were in the field by 25 September (Appendix 1 contains lists of the assessment teams and the areas covered by them).

The brief given to the teams was:

• to report on the immediate food and water supplies in rural communities;
• to identify places in critical need, and to assess the probable food and water situation until the end of 1997 and the early part of 1998;
• to assess whether people were moving out of areas in response to a lack of food or water;
• to assess local drinking water supplies;
• to assess the current health situation and to look in particular for evidence of severe malnutrition among young children;
• and to report on the existence and the state of local area services and communications.

It was stressed to the teams that, wherever possible, reports should be based upon direct personal observations, and should cover a range of indicators which should include the examination of gardens, women’s carrying bags, cook houses and dwelling houses, roadside and urban market places, the cash earning abilities of people and their cash reserves and how much long cash income would continue to be available; their management of domesticated animals and their use of famine foods. Observations of aid posts and health centers were also required. Water supplies were to be examined where possible and drinking water on hand be examined. In many locations, discussions were held with groups of women in separate meetings.

The assessment also provided the opportunity to trial a method of on-going monitoring and assessment to be used over the next 12 months.

A form was used in order to collect this information systematically from a large number of observers over a range of administrative units down to the level of individual census units (villages). The form was prepared in Canberra and amended after consultations with Papua New Guinean participants (Appendix 2).

The assessment covered all of the key steps recommended in the World Health Organisation Emergency Preparedness and Response (1990) guidelines for Rapid Health Assessment in Suspected Famine Situations (ERO/EPR/90.1.8), although this document was not available prior to the assessment being undertaken.

The teams were assigned areas, provinces or parts of provinces and were instructed wherever possible to base their reports on direct personal observations. Cash advances were made to team leaders to cover transport and accommodation costs.

Copies of the forms used by the teams were sent to all Provincial administrations and they were invited to make their own observations based on questions contained in the forms. The purpose of doing this was to try and reduce major differences between assessments made by the Rapid Assessment Teams and local authorities, and to improve the quality of provincial based reporting. The last named tended to be generalised and uncritical, and to be of little use in detailed planning or costing.

Prior to leaving Port Moresby, Brian Deutrom and staff in the Department of Provincial and Local Government Affairs (DPLGA) led by Mr Kanji Ukegawa, a JICA volunteer, assisted by Mr Gaina Bere, began to set up Microsoft Access database files to receive information from the forms as they were faxed or phoned in from teams in the field.

The National Disaster and Emergency Centre provided operational support to the field teams for 16 hours per day and received and passed on messages and information to team leaders.

The assessment was completed by 11 October 1997, when the last of teams arrived back in Port Moresby. From 8 October to 14 October 1997, information from 638 forms was entered into the database. Staff of the DPLGA worked with the Australian consultants for long hours through the weekend.

A great many people assisted with the assessment, first and foremost the men and women of the rural areas who gave freely of their time and knowledge. Organisations which provided direct assistance are: Department of Provincial and Local Government Affairs, Department of Agriculture and Livestock,, the Australian High Commission, Provincial administrations, Porgera Joint Venture, Ok Tedi Mining Ltd, Chevron, Melanesian Tours and the Australian National University. We worked closely with the CARE Australia team of Graham Henderson (health), David Hill (logistics), and Gerard Jacobson (water supply) and their observations and ideas are reflected in this report.