Tag: selection bias

In a recent article in GRL, Emerging selection bias in large-scale climate change simulations, Kyle L. Swanson describes how experiments might go wrong as we seek a desired solution. He hypothesise a common wish to reproduce the recent warming in the Arctic, has lead to less diversity among models with convergence towards some common solution.

He writes the current generation ensembles of model simulations are statistically inconsistent with the observed shifts in in both the mean surface air temperature as well as the frequency of  extreme monthly mean temperature events due to climate warming, despite a marked reduction in the spread among ensemble members that by itself suggests convergence towards some common solution. This convergence indicates the possibility of a selection bias based upon warming rate.

I particularly liked his example of what happened after Robert A. Millikan’s original measurement of the charge of the electron: “Millikan’s original measurement was slightly erroneous due to the use of an incorrect value of the viscosity of air. In the decades following Millikan’s work and his subsequent Nobel Prize, other investigators empirically measured the electron charge. When they got a number that was too high above Millikan’s, they thought something must be wrong–and they would look for and find a reason why something might be wrong. When they got a number close to Millikan’s value they didn’t look so hard. And so they eliminated the numbers that were too far off, and did other things like that.”

In our paper in Malaria Journal, A dynamic model of some malaria-transmitting anopheline mosquitoes of the Afrotropical region. I. Model description and sensitivity analysis, we touch upon this issue by stating “A model is mental copy that describes one possible representation of a system… We present an alternative formulation of the dynamics of An. gambiae s.s. and An. arabiensis…  if malaria modelers move towards the ensemble thinking widely adopted in the climate community, this model could be one representation of historical and future changes for malaria. The aim of such an ensemble would be to deal with uncertainties in the system. Ultimately, the goal would be to produce policy-relevant information including uncertainty.”

The basic principles of climate are quite easy to understand; the sun is heating the Earth, with more radiation hitting the equator compared to the poles. Seasons are caused by the tilt of the Earth’s rotational axis away or toward the sun as it travels through its year-long path around the sun. From there you can add clouds, water, ice, land masses, aerosols, precipitation, and even mountain and building shadows if you are interested in local weather. The dynamics of malaria can be simplified in the same manner; an uninfected mosquito bites an infectious human, lives for about ten days, and from that time it can infect new humans every time it bites a human. From this simple model it has been demonstrated that killing mosquitoes is an efficient way of reducing malaria, but it is not able to tell us how efficient. To get a realistic representation of how emissions of greenhouse gasses influence climate, or how malaria is reduced by distributing bed nets we need more complex models together with observations. As we add more realism to the system, we also parameters with uncertainty, uncertainty which was there from the beginning, but which becomes visible as we describe them. If we intended to use an ensemble of models to estimate the effect of for example bed nets against malaria, we could get a very precise estimate if all models were identical, or measured their success by managing to keep malaria away from South Africa (many malaria models predict malaria is widespread there). There are still many things we do not know about malaria, and I hope modellers of malaria do not do the same mistake it seems like global climate models are moving towards; a common solution due to wrong reasons.