Technical efficiency in relation to climate change factors The evaluation of technical efficiencies provides opportunities to identify

aspects of the farming systems that need to be changed to increase cost effectiveness of the production. However, the technical efficiencies have not been determined for most different types of aquaculture systems . For example, only the technical efficiencies of trout pond farming in the Turkish Black Sea Region (Cinemre, 2006), tilapia pond operations in the Philippines (Dey et al., 2004), semi-intensive/intensive carp farming in India (Vereena et al., 1999; Sharma & Leung 2000), Chinese fish farms polyculture (Sharma, 1999) and the carp farming of Bangladesh (Ferdous and Murshed-e-Jahan, 2008) are examined. If changes have to be made in the aquaculture systems to combat climate change, the evaluation of the technical efficiency enables an assessment whether these changes permit farms to remain productive and cost effective or not.

The technical efficiency of pangasius farms in the Mekong Delta of Vietnam was on par with fish farms operated in other countries (Chapter 4). Pangasius farmers in the flood prone areas of the upstream and mid-stream regions had a larger scale of operation, whereas salinity intrusion reduced the scale of farms located in the downstream region (Chapter 4). More efficient use of inputs, such as feed, seed and labour, and an increased scale of operation, as well as intensified training will be essential for lifting productivity. The overall adaptation strategy (e.g. stronger dykes) requires a one-time investment, while the increased salinity level downstream will require also increased recurrent operational costs either at sector level (e.g. breeding of salinity tolerant pangasius) or at farm level (e.g. a prolonged nursery phase or reduced number of stockings/harvests).

Our analysis, however, did not establish a firm influence of climate change on the technical efficiency. Perhaps, data sets of a larger number of farms on a longer timeframe that include years/crops that were affected by rare weather events, are required to obtain more robust and statistically valid observations.

Moreover these longitudinal datasets need to include one or more specific parameters related to climate change impacts such as salinity level or/ and height and length of water (flood) level.

6.4. Adaptation options

According to Smit et al. (2000), adaptive responses to climate change varies in process and forms. Carter et al. (1994), distinguished between ‘autonomous’ (i.e. automatic, spontaneous, passive or natural) adaptation(s) and ‘planned’ (i.e. strategic or active) adaptation(s). Smit et al. (2000) noted that the autonomous adaptations in unmanaged natural systems diffed from consciously planned adaptations based on intent or purposefulness with respect to an expected climate stimulus. Adaptations that are initiated by public agencies, are usually planned but adaptations by individuals or groups may be autonomous or planned, or some combination of the two (Smit et al., 2000). IPCC (2014) indicated several available measures to help aquaculture to adapt to climate change impacts. For example, the shellfish farmers in the north-western US adapted autonomously to changes in the acidity of seawater by immediately blocking the intake when pH levels fall below a certain threshold, or actively planned to move their hatcheries to Hawaii. To adapt to the expected climate change impacts the proposed and already applied measures by pangasius farmers included increasing dyke height, and, especially downstream, decreasing the number of stockings or stocking larger fish. Larger pangasius are probably capable of tolerating higher salinity levels without a negative effect on their performance (Chapter 3), but conclusive scientific eidence is yet to obtained.

Facing the risk of flooding and salinity intrusion induced by sea level rise on the pangasius farming sector in the Mekong Delta (Chapter 2, 3 and 4), Chapter 5 proposed plausible measures that include autonomous and planned adaptation(s). The farmers in different administrative units are or feel affected differently by the impacts of climate change. Our straightforward decision support tree aims at helping the stakeholders by clearly defining the conditions under which specific decisions should be taken. The proposed suitable autonomous adaptive measures of pangasius farmers are: increasing

pond dyke height, changing pond practice(s) (spontaneous adaptation) and shifting to farm other (more salinity tolerant) species (spontaneous adaptation). However, the last adaptation measure may only be selected by few farmers because it requires investments in new buildings, other professional networks and restructuring of the ponds (Chapter 5). The planned adaptation measures include improving flood protection dykes (i.e. strategic adaptation), developing a salinity tolerant strain of pangasius (active adaptation), and stocking this salinity tolerant strain by the pangasius farmers (active adaptation). These measures should be implemented by the national or provincial governments, public agencies or private companies.

Chapter 5 revealed the costs of autonomous adaptation. When pangasius farmers improve their pond dyke by increasing the height, they increase the total variable costs (ha-1 _crop-1_{) by about 0.34% and 0.25% in the up- and mid-} stream regions, and in the downstream region, respectively. The improvement of flood protection dykes reduces the need for further autonomous adaptations of individual pangasius farmers and is also beneficial for other sectors but may have negative effects for biodiversity and rice-farmers (Lebel et al, 2009; Marchand et al. 2014).

When balancing the evidence, taking into account socio-economic conditions, costs and benefits, the most convenient option for the downstream farms would be to develop a salinity tolerant strain of pangasius. This option likely results in the least disruption to livelihoods, requires minimal changes in infrastructure and permits to maintain the well-established market channels. In the wake of the current advances in molecular genetic techniques applicable for selective breeding, innovative technologies may allow to achieve desired targets relatively soon. Chapter 5 estimated the annual cost of such a breeding program to produce and maintain a salinity tolerant strain of pangasius at about US$ 120,000. However, as pangasius is a late maturing species, this type of multi-generation breeding requires a long term program of selection and breeding to maintain the desired characteristics in the produced broodstock, and cannot be dealt with through a project approach as most governments tend to do (Chapter 5). This means a 0.4% increase of present production cost for the downstream farms producing about 10% of the export volume. Either a larger private sector company/organisation or a

farmers organisation should take a lead role to develop and supply the salinity tolerant pangasius seed.