Evaluating the potential for statistical decadal predictions of sea surface temperatures with a perfect model approach

We explore the potential for making statistical decadal predictions of sea surface temperatures (SSTs) in a perfect model analysis, with a focus on the Atlantic basin. Various statistical methods (Lagged correlations, Linear Inverse Modelling and Constructed Analogue) are found to have significant skill in predicting the internal variability of Atlantic SSTs for up to a decade ahead in control integrations of two different global climate models (GCMs), namely HadCM3 and HadGEM1. Statistical methods which consider non-local information tend to perform best, but which is the most successful statistical method depends on the region considered, GCM data used and prediction lead time. However, the Constructed Analogue method tends to have the highest skill at longer lead times. Importantly, the regions of greatest prediction skill can be very different to regions identified as potentially predictable from variance explained arguments. This finding suggests that significant local decadal variability is not necessarily a prerequisite for skillful decadal predictions, and that the statistical methods are capturing some of the dynamics of low-frequency SST evolution. In particular, using data from HadGEM1, significant skill at lead times of 6–10 years is found in the tropical North Atlantic, a region with relatively little decadal variability compared to interannual variability. This skill appears to come from reconstructing the SSTs in the far north Atlantic, suggesting that the more northern latitudes are optimal for SST observations to improve predictions. We additionally explore whether adding sub-surface temperature data improves these decadal statistical predictions, and find that, again, it depends on the region, prediction lead time and GCM data used. Overall, we argue that the estimated prediction skill motivates the further development of statistical decadal predictions of SSTs as a benchmark for current and future GCM-based decadal climate predictions.

Comparing the UK Met Office climate prediction system DePreSys with idealized predictability in the HadCM3 model

The initial condition effect on climate prediction skill over a 2-year hindcast time-scale has been assessed from ensemble HadCM3 climate model runs using anomaly initialization over the period 1990–2001, and making comparisons with runs without initialization (equivalent to climatological conditions), and to anomaly persistence. It is shown that the assimilation improves the prediction skill in the first year globally, and in a number of limited areas out into the second year. Skill in hindcasting surface air temperature anomalies is most marked over ocean areas, and is coincident with areas of high sea surface temperature and ocean heat content skill. Skill improvement over land areas is much more limited but is still detectable in some cases. We found little difference in the skill of hindcasts using three different sets of ocean initial conditions, and we obtained the best results by combining these to form a grand ensemble hindcast set. Results are also compared with the idealized predictability studies of Collins (Clim. Dynam. 2002; 19: 671–692), which used the same model. The maximum lead time for which initialization gives enhanced skill over runs without initialization varies in different regions but is very similar to lead times found in the idealized studies, therefore strongly supporting the process representation in the model as well as its use for operational predictions. The limited 12-year period of the study, however, means that the regional details of model skill should probably be further assessed under a wider range of observational conditions.

Effective communication to improve udder health: can social science help?

Improved udder health requires consistent application of appropriate management practices by those involved in managing dairy herds and the milking process. Designing effective communication requires that we understand why dairy herd managers behave in the way they do and also how the means of communication can be used both to inform and to influence. Social sciences- ranging from economics to anthropology - have been used to shed light on the behaviour of those who manage farm animals. Communication science tells us that influencing behaviour is not simply a question of „getting the message across‟ but of addressing the complex of factors that influence an individual‟s behavioural decisions. A review of recent studies in the animal health literature shows that different social science frameworks and methodologies offer complementary insights into livestock managers‟ behaviour but that the diversity of conceptual and methodological frameworks presents a challenge for animal health practitioners and policy makers who seek to make sense of the findings – and for researchers looking for helpful starting points. Data from a recent study in England illustrate the potential of „home-made‟ conceptual frameworks to help unravel the complexity of farmer behaviour. At the same time, though, the data indicate the difficulties facing those designing communication strategies in a context where farmers believe strongly that they are already doing all they can reasonably be expected to do to minimise animal health risks.

Farmers' management of rice varietal diversity in the mid-hills of Nepal: implications for on-farm conservation and crop improvement

Season-long monitoring of on-farm rice (Oryza sativa, L.) plots in Nepal explored farmers' decision-making process on the deployment of varieties to agroecosystems, application of production inputs to varieties, agronomic practices and relationship between economic return and area planted per variety. Farmers deploy varieties [landraces (LRs) and modern varieties (MVs)] to agroecosystems based on their understanding of characteristics of varieties and agroecosystems, and the interaction between them. In marginal growing conditions, LRs can compete with MVs. Within an agroecosystem, economic return and area planted to varieties have positive relationship, but this is not so between agroecosystems. LRs are very diverse on agronomic and economic traits; therefore, they cannot be rejected a priori as inferior materials without proper evaluation. LRs have to be evaluated for useful traits and utilized in breeding programmes to generate farmer-preferred materials for marginal environments and for their conservation on-farm.