SEASONAL FORECAST SUMMARY UPDATE. Nov 13, 2014

Transcription

1 SEASONAL FORECAST SUMMARY UPDATE Nov 13, 2014

2 Contents Executive summary of the strongest seasonal signals... 1 How good were last season s forecasts?... 2 Introduction... 4 Global view... 5 Global outlook... 5 Global storm outlook... 6 Regional focus... 6 Japan... 6 Australia... 7 United States... 7 Climate signals... 8 Seasonal climate forecast skill... 9 Information sources... 9 For further information contact: Dr. James Done Project Scientist and Willis Research Fellow National Center for Atmospheric Research Earth System Laboratory done@ucar.edu Website: P.O. Box 3000 Boulder, CO USA Geoffrey Saville Willis Research Network Atmospheric Hub Leader Willis Global Analytics and WRN geoffrey.saville@willis.com Phone: +44 (0) Lime Street London, EC3M 7DQ UK

3 Executive summary of the strongest seasonal signals A review of seasonal climate forecasts issued by major international forecast centers suggests the strongest seasonal signals for the boreal winter include: U.S.: Increased odds of warm temperatures over western and northern regions and Alaska while southern regions have a signal for cool temperatures. Increased odds of wet conditions across southern States and up the Eastern seaboard. Australia: Increased odds of warm temperatures for most regions and increased odds of dry conditions over northern and eastern regions. Japan: Warmer-than-average temperatures are favored over southern regions with increased odds of wet conditions over south facing regions of the major Islands. Some north coast prefectures have an increased chance of less than normal snowfall. Precipitation and temperature signals are summarized below: Table 1 Latest seasonal precipitation signals (Dec-Jan-Feb 2014/2015) Country Region Japan south-facing regions Australia east and north U.S. south and east U.S. southern Alaska U.S. northern Rockies U.S. Great Lakes Most likely seasonal precipitation category (vs. normal) Table 2 Latest seasonal temperature signals (Dec-Jan-Feb 2014/2015) Country Region Australia Japan - south U.S. west and north U.S. Alaska U.S. south Most likely seasonal temperature category (vs. normal) Page 1

4 How good were last season s forecasts? Verifying probabilistic seasonal forecasts is not as simple as verifying a weather forecast of, say, 5 C in London tomorrow because a probabilistic seasonal forecast is for all conditions to occur (cold, normal and hot temperatures, for example) each with an assigned likelihood. Verification of seasonal forecasts is an ongoing science that aims to capture and easily communicate multiple aspects of the forecast including reliability, discrimination, bias, success rate and value (IRI 1 ). However, we restrict ourselves in this document to presenting the forecasts alongside the actual outcome. The main signals from last season s August-September-October forecasts (WRN report issued on the 31 st of July 2014) are summarized for precipitation and temperature in Tables 3 and 4 (respectively) below: Table 2 - Precipitation signals from previous report, July Table 4 - Temperature signals from previous report, July Country Region Most likely seasonal precipitation category Country Region Most likely seasonal temperature category (vs. normal) (vs. normal) U.S. Intermountain West U.S. West Australia Northeast U.S. Southeast Australia Southeast U.S. Alaska U.S. Northern Plains Australia Northeast Australia Southeast Australia Southwest For the main focus regions of the U.S. and Australia, the forecasts verified reasonably well when comparing the forecast category of greatest likelihood (i.e., below-normal, normal and above-normal) with the observed category. For Australia, the drier than normal conditions were captured in the southeast (Figure 1) but the drier than normal conditions in the middle and North of the country were largely missed. The temperature forecast fared much better with warmer than normal conditions captured in southern regions (Figure 1) but the forecast for warm conditions in the northeast failed to materialize. Figure 1: Actual mean temperature (left) and rainfall (right) for the period Aug-Sept-Oct 2014 compared to normal. Source: BOM 4. Page 2

5 The forecast for the U.S. fared much better. The broad pattern of warmer than normal temperatures in Western and southeastern regions was well forecast (Figure 2). The cooler than normal conditions over the Northern Plains was also well forecast. Wetter than normal conditions over the Intermountain West were well captured (Figure 2). Figure 2: Actual mean temperature (left) and precipitation (right) for the period Aug-Sept-Oct 2014 compared to normal. Source: NOAA Climate Prediction Center 5. Page 3

6 Introduction This document provides a summary of seasonal climate forecasts for boreal winter 2014/2015 tailored to the needs of the reinsurance industry. The strongest temperature and rainfall signals are identified across readily available seasonal forecasts issued by the major international forecast centers and signals are interpreted in the context of key hazard parameters such as levels of storm activity together with a discussion of potential industry impacts. Although this document takes a global view, regions of greatest exposure are emphasized with particular focus on the U.S., Australia and Japan.... regions of greatest exposure are emphasized with particular focus on U.S., Australia and Japan. Detailed background information on the science, production and interpretation of seasonal climate forecasts are provided in the accompanying Seasonal Forecast Summary: Go-To Guide, but three points are summarized briefly here to aid interpretation: 1. Official seasonal climate forecasts are a synthesis of a diverse range of forecast information from dynamical and statistical models together with objective (and sometimes subjective) weighting. A majority of forecast information is generated in-house, but it is common for forecast centers to pool information sources. This can result in a degree of commonality between forecasts from different centers. 2. Seasonal forecasts (months to seasons ahead) are probabilistic, so they must be interpreted differently to interpretations of conventional weather forecasts (forecasts of just a few days ahead). We should not expect only one outcome to occur, but rather multiple potential outcomes. For example, a seasonal forecast of increased likelihood of above-normal temperature should not be interpreted as we expect a warmer than normal season but rather we expect increased likelihood of a warmer than normal season and reduced likelihood of a normal season and reduced likelihood of a cooler than normal season. Without this probabilistic way of thinking, other climatic conditions that might still occur can be overlooked, and responses can be inappropriate. 3. Forecasts can change substantially from one issuing month to the next, primarily due to new information coming from the dynamical models. Page 4

7 Global view Global outlook Global seasonal forecasts are produced by many centers. Two of the most comprehensive are discussed here. Forecasts produced by the International Research Institute for Climate and Society (IRI) are primarily a synthesis of forecasts from many dynamical models that respond to expected ocean temperature patterns. Each model forecast is weighted according to seasonally and regionally dependent skill based on past forecasts and may be adjusted for known biases or late breaking climate information. The Asia Pacific Economic Cooperation Climate Center (APCC), another rich source of global seasonal climate forecast information, produces forecasts primarily based on a large number of dynamical model information. IRI s global temperature forecast issued in October for December-January-February (Figure 3) indicates increased likelihood of above-normal temperature over nearly all countries with the strongest signals over Africa, South America, and Southeast Asia. The only signal for increased likelihood of below-normal temperatures appears over the southern U.S. This overall warm signal is likely due to the continued expectation of El Niño conditions becoming established this boreal winter. Figure 3: Global temperature (left) and rainfall (right) forecast for December-January-February 2014/2015. Source: IRI 1. IRI s global rainfall forecast for December-January-February shows signals for increased likelihood of below-normal rainfall over northern South America, Indonesia and parts of the northern U.S. (Figure 3). The only coherent signal for increased likelihood of above-normal rainfall lies across southern regions of the U.S. and coastal Alaska. APCC is in general agreement with IRI on these global signals. These rainfall patterns are again largely due to a continued expectation of El Niño conditions becoming established later this winter. As we enter boreal spring, these temperature and rainfall signals largely persist. Page 5

8 Global storm outlook Dynamical models not only forecast temperature and rainfall but also forecast the full state of the climate, meaning we can explore many other climate variables. Of greatest relevance to industry are indicators of unusual levels of extratropical cyclone storm activity. One such indicator is the geopotential height at 500mb, which is the height above the Earth s surface at which the atmospheric pressure is 500mb. Low values indicate cooler and stormier weather than normal and high values indicate warmer and more settled weather than normal. Figure 4 shows the APCC forecast departure from normal geopotential height at 500mb for November-December-January 2014/2015. More settled weather than normal extends over Japan and the northeast U.S. Weak signals for more unsettled weather than normal exists over the Western British Isles, Coastal Alaska and New Zealand. Figure 4: Forecast departure from normal geopotential height at 500mb for November-December-January 2014/2015. Cool colors indicate stormier weather than normal and warm colors indicate more settled weather than normal. Source: APCC 2. Regional focus Japan The Japan Meteorological Agency issues detailed seasonal forecasts for Japan. The forecast issued on Oct 25, 2014 for December-January-February 2014/2015 indicates abovenormal temperatures for central and southern regions (Figure 5). The forecast for rainfall shows increased likelihood of above normal rainfall along the south facing regions of the major Islands (Figure 5). The temperature outlook is broadly in line with the global outlook but the rainfall outlook is rather different with climatological conditions expected by IRI (Figure 3) and even a drier than normal signal by APCC (not shown). Figure 5: Forecast temperature (top) and rainfall (bottom) for the period December-January-February 2014/2015. Source: Japan Meteorological Agency 3. Page 6

9 Another seasonal product from the Japan Meteorological Agency is the seasonal snowfall forecast. The outlook (Figure 6) shows for many parts of north coastal prefectures to receive normal amounts of snowfall, although further south along the northern parts of Kansai and Chūgoku there is a slightly increased likelihood of seeing less snowfall than normal. Elsewhere there is no strong signal present. Figure 6: Forecast snowfall for the period December- January-February 2014/2015. Source: Japan Meteorological Agency 3. Australia The Bureau of Meteorology (BOM: Australia s national weather, climate and water agency) produces detailed seasonal climate forecasts for the Australia region, primarily based on a large number of dynamical model forecasts. The Bureau forecast issued on Oct 30, 2014 for the period November-December-January indicates warmer than normal conditions are more likely over most regions with the warmest signal over northern and eastern regions (Figure 7). This has some support from the global outlooks. Drier than normal conditions are most likely over northern and eastern regions with the driest signal over Queensland (Figure 7). Figure 7: Forecast likelihood of exceeding (left) the median maximum temperature and (right) the median rainfall, for the period November-December-January 2014/2015. Source BOM 4. United States Our richest source of seasonal climate information for the U.S. is provided by The National Oceanic and Atmospheric Administration s (NOAA s) Climate Prediction Center (CPC). NOAA s approach consolidates a diverse range of statistical and dynamical seasonal forecasts in a manner that favors models that have performed well in past seasons. There is a tendency to rely on dynamical model guidance for the first few months, transitioning to statistical models for the later period to reflect the relative skill of each approach with lead-time. Outlooks are issued around the middle of every month. This report summarizes the forecast issued Oct 16, 2014 for the period Dec-Jan-Feb 2014/2015. Perhaps unsurprisingly, NOAA s seasonal forecasts for the U.S. are broadly consistent with the global views since they share some data. Increased likelihood of above-normal temperature covers western and northern regions including Alaska, with the strongest signals over Washington State, Oregon and California (Figure 8). An increased likelihood of below-normal temperatures is predicted over southern states. Page 7

10 Increased likelihood for below-normal rainfall lies over the Northern Rockies and the Great Lakes region while a signal for above normal rainfall extends across the southern U.S. (including southern California) and up the Eastern seaboard, and coastal Alaska (Figure 8). Figure 8: Forecast issued Oct 16, 2014 for U.S. temperature (left) and rainfall (right) for the period Dec-Jan-Feb 2014/2015. Source: NOAA Climate Prediction Center 5. Climate signals The El Niño-Southern Oscillation (ENSO), our biggest source of seasonal forecast skill, remains in a neutral phase, despite continued forecasts throughout 2014 for the emergence of El Niño. The potential for a weak El Niño to become established in boreal winter remains high; latest indications are lining up and point towards the process being underway. Planning should certainly take this potential into account. For much of 2014 we have experienced warm ocean conditions across the equatorial Pacific but the atmosphere has been reluctant to come into line with an El Niño response. However, there is now evidence of a reversal of winds over the equatorial Pacific, which, if sustained through November, could tip the system towards El Niño. The consensus forecast from IRI has a 65% probability of El Niño forming by December with only a 35% probability of neutral conditions and zero probability of La Niña. Much of the seasonal forecast signals are predicated on weak El Niño conditions becoming established, although a low-end moderate El Niño or ENSO neutral conditions are also possible. In addition to ENSO and its global teleconnections, anomalies in mid-latitude ocean temperatures are also considered in the seasonal forecasts. Currently, warmer than average ocean temperatures lie across the North Pacific and the western North Atlantic and contribute to the warm and wet signals for southern Alaska and the far northeast U.S. The expected El Niño dominates the signals for the rest of the U.S. including the warm and wet signals over southern California and cool and wet signals over Southern States associated with an amplified storm track. The warm and dry signal over northern U.S. is associated with the Polar Jetstream located further north than normal typical under El Niño conditions. Much of Australia has a signal for dry and warm conditions reflecting warmer than average ocean temperatures across the equatorial Pacific. Even if a fully-fledged El Niño doesn t develop, these ocean temperature patterns are still likely to have a warming and drying effect over Australia. The Indian Ocean Dipole remains neutral and has little influence on Australian climate during the Austral summer. Page 8

11 Seasonal climate forecast skill Our experience with daily weather forecasts has taught us that the actual weather can sometimes be quite different from the forecast. The same is true for seasonal climate forecasts, but are they accurate enough to be useful? IRI provides a wealth of forecast skill information for different lead times, seasonal and regions. Perhaps the skill score of most relevance to industry is the rate of return. The rate of return tells us how much money would be made if one invested based on the forecast with odds corresponding to the probabilities given in the forecast. The notional amount of money made is therefore related closely to the accuracy of the forecast. Figure 9: Rate of return for Dec-Jan-Feb temperature at 0.5 months lead-time. Gray indicates no (or negative) skill and purple indicates highest skill. Source: IRI 1. Figure 9 shows rates of return for Dec-Jan-Feb temperature forecasts at 0.5 months lead-time. Highest rates of return are over tropical regions with moderate values also over northern Australia, and western and southern U.S. For precipitation, rates of return are typically much lower (not shown) with values almost universally less than 0.1 and some values dipping below zero. This reflects the lower predictability for rainfall compared to temperature, due to rainfall resulting from more complex climate processes. Owing to the peculiarities of any single forecast, benefits will increase with the number of seasons. Information sources 1 The International Research Institute for Climate and Society. 2 The Asia Pacific Economic Cooperation Climate Center. 3 Japan Meteorological Agency. 4 The Bureau of Meteorology. 5 The National Oceanic and Atmospheric Administration s Climate Prediction Center. Page 9

12 How can we help? To find out how we can offer you an extra depth of service combined with extra flexibility, simply contact us. Begin by visiting our website at or calling your local office. Copyright 2014 Willis Limited / Willis Re Inc. All rights reserved: No part of this publication may be reproduced, disseminated, distributed, stored in a retrieval system, transmitted or otherwise transferred in any form or by any means, whether electronic, mechanical, photocopying, recording, or otherwise, without the permission of Willis Limited / Willis Re Inc. Some information contained in this document may be compiled from third party sources and we do not guarantee and are not responsible for the accuracy of such. This document is for general information only and is not intended to be relied upon. Any action based on or in connection with anything contained herein should be taken only after obtaining specific advice from independent professional advisors of your choice. The views expressed in this document are not necessarily those of Willis Limited / Willis Re Inc., its parent companies, sister companies, subsidiaries or affiliates (hereinafter Willis ). Willis is not responsible for the accuracy or completeness of the contents herein and expressly disclaims any responsibility or liability for the reader's application of any of the contents herein to any analysis or other matter, or for any results or conclusions based upon, arising from or in connection with the contents herein, nor do the contents herein guarantee, and should not be construed to guarantee, any particular result or outcome. Willis accepts no responsibility for the content or quality of any third party websites to which we refer. The contents herein are provided for informational purposes only and do not constitute and should not be construed as professional advice. Any and all examples used herein are for illustrative purposes only, are purely hypothetical in nature, and offered merely to describe concepts or ideas. They are not offered as solutions to produce specific results and are not to be relied upon. The reader is cautioned to consult independent professional advisors of his/her choice and formulate independent conclusions and opinions regarding the subject matter discussed herein. Willis is not responsible for the accuracy or completeness of the contents herein and expressly disclaims any responsibility or liability for the reader's application of any of the contents herein to any analysis or other matter, nor do the contents herein guarantee, and should not be construed to guarantee, any particular result or outcome.