The heat is on Peter Guttorp peter guttorpnr

The heat is on! Peter Guttorp peter. guttorp@nr. no guttorp@uw. edu http: //www. stat. washington. edu/peter

The greenhouse effect Heat comes in from the sun Shortwave radiation Earth gets warmed up by the heat Earth radiates heat back Longwave radiation Greenhouse gases absorb much energy in radiating heat Atmosphere warms (15°C instead of -18°C) Main greenhouse gases: Water vapor Carbon dioxide Methane

The greenhouse effect Joseph Fourier (1768 -1830) realized that Earth ought to be a lot cooler than it is. John Tyndall (1820 -1893) found that water vapor and CO 2 are greenhouse gases Svante Arrhenius (1859 -1927) calculated how changes in CO 2 can heat the planet

What is climate? Climate is what you expect; weather is what you get. Heinlein: Notebooks of Lazarus Long (1978)

Outline Measurements Models Local impact Projections

Measurements

Measuring global surface temperature

Homogenization summertime correction miscalibrated thermometer screen painted white? urbanization

Global temperature measurements Marine data

Comparison between estimates

Is there a trend? An Ac t of Dog Global temperature

Models

Climate modeling

The issue of gridding Hurricanes Clouds Glaciers

Comparing global climate models to data

30 -year distributions

Local impact

Comparing climate model output to weather data Global models are very coarse Regional models are driven by boundary conditions given by global model runs In either case, describes distribution of weather, not actual weather Consider a regional model driven by “actual weather” Stockholm 50 km x 50 km grid, 3 hr resolution (SMHI-RCA 3; ERA 40)

Stockholm data issues Location was moved twice (1875, 1960) Calibration (1826: 0 reads as +0. 75; 1858, 1915; annual thereafter)

How well does the climate model reproduce data?

Model problem? Annual average temperature over the grid square containing the Stockholm site is about 1. 7°C warmer than the observed average Model calculates separately open air, forest, and water/ice. Do we need finer resolution?

Open air predictions Using 12. 5 km version of RCA 3, forced by ERA 40, looking at only open air predictions (77% of grid square is open air)

Is the station really in open air?

Comparison to forested model output

Projections

Why not predictions? Climate models need input of greenhouse gases, solar radiation, land use etc. To use climate models for prediction, must predict also these input variables. Instead, set up scenarios (reasonable values of the input variables). Run models with these inputs. We call that projections.

Projecting sea level rise Sea levels rise due to • warming of oceans • melting of land ice Most climate models do not output sea level Strategy: relate global mean temperature to global mean sea level relate global to local sea level Use projections of temperature to project local sea level

Bergen Cultural Heritage Site Storm surges up to 1. 4 m Land rise 2. 6 mm/year

Projections

Components of uncertainty

Using uncertainty in decision making Do Bergen authorities need to address sea level rise? If so, when? Adaptation costs: Outer barrier 30 B NOK (5 B CAD) Inner barriers 1. 1 B (0. 2 B) Need cumulative storm surge damage costs.

Current storm surge damage costs

Change due to sea level rise

Simulate damages Draw random annual cost Draw random increase factor path Draw random sea level path Accumulate costs over time Look at upper 95 th percentile of cumulative costs

When is an adaptation measure beneficial? Outer barrier Inner barriers

Some references P. Guttorp and J. Xiu (2011): Climate change, trends in extremes, and model assessment for a long temperature time series from Sweden. Environmetrics 22: 456 -463. P. F. Craigmile and P. Guttorp (2013): Can a regional climate model reproduce observed extreme temperatures? Statistica 73: 103 -122. P. Guttorp (2014): Statistics and Climate. Annual Reviews of Statistics and its Applications 1: 87 -101. P. Guttorp, D. Bolin, A. Januzzi, D. Jones, M. Novak, H. Podschwit, L. Richardson, A. Särkkä, C. Sowder and A Zimmerman (2014): Assessing the uncertainty in projecting local mean sea level from global temperature. Journal of Applied Meteorology and Climatology 53: 2163 -2170.

Uncertainty in cumulative damage