National Weather Service-Pennsylvania State University Weather Events New Zealand heavy rain and flood event-draft by Richard H. Grumm National Weather Service State College PA 16803 Abstract:. A record rain event, reported as one of the largest events in decades, brought flooding and landslides to the South Island. The heavier rainfall was confined to the northern and western side of the Island which was exposed to the strong northwesterly flow into the Island. A deep plume of above normal precipitable water moved into the Island from the north. The deep northwesterly flow about a strong mid-tropospheric ridge to the southeast and strong trough to the northwest brought the deep tropical moisture plume into the Island. Using re-analysis climate data (R-Climate) this event clearly illustrates the value of climate data to diagnose and predicted record and near record heavy rainfall events. Model climate (M-Climate) based forecast are presented to show the value of internal model climate data to aid in predicting extreme weather events.
1. INTRODUCTION A strong subtropical ridge and deep trough (Fig. 1) brought deep southerly flow and plume, or atmospheric river (AR: Neiman et al. 2008;Carlson 1980), of deep tropical moisture into New Zealand (Fig. 2). The deep plume of moisture, combined with strong low-level winds and the terrain of New Zealand, produced a high impact heavy rainfall event over the South Island. The heavy rainfall, reported as the worse rain event in decades (BBC 2011), caused the evacuation of over 100 people as a state of emergency was declared due to landslides and swollen rivers, damaged roads and bridges and cut-off towns and villages. The heavy rains, mainly in the Tasman, caused floods and damaged numerous homes. Over 500 mm of rainfall was observed near Takaka. The heavier rainfall amounts were confined to the northern and western side of the South Island 1. The Tasman region is on the northeastern edge of the main South Island and includes the towns of Richmond, Motueka, Collingwood (Fig. 3), Takaka, and Brightwater. Similar to many record high impact weather events (HIWE), this event is diagnosed using reanalysis climate (R-Climate) data as described by Hart and Grumm (2001) and Graham and Grumm (2010). R-Climate data often aids in depicting features which depart significantly from normal. These concepts have been applied to output from numerical weather prediction models and ensembles (Junker et. al 2008) and appear to aid in predicting synoptically forced high impact weather events. The concept of using an internal model or ensemble climate (M-Climate) has been used by the European Center for Medium-Range Weather Forecasting (ECMWF). An extreme forecast index (EFI) has been developed (Lalaurette 2003) which aids in identifying when the ensemble forecast system is predicting an extreme event relative to the internal model climatology.. This paper will document the pattern and the value of R-Climate data in diagnosing the New Zealand heavy rainfall event of 13-14 December 2011 2. The focus is on the R-Climate based standardized anomalies as a tool to both analyze and predict this and similar events. The forecast section of this paper will show the value of using climate data and internal ensemble prediction system climate data to better anticipate and predict extreme weather events. The concepts of both R-Climate and M-Climate are presented here to show the value of and the complimentary role R-Climate and M-Climate data play when forecasting of extreme events. 2. Methods and Data The NCEP GFS is used to re-produce the conditions associated with the storm to include the large scale pattern. The standardized anomalies are displayed in standard deviations from normal as in Hart and Grumm (2001) and are computed using the climatology from the NCEP/NCAR global reanalysis data (Kalnay et al. 1996). The focus 1 Information from Weather Watch New Zealand http://weatherwatch.co.nz/ 2 All times are referenced in UTC time and do not reflect the local time in New Zealand which was 14-15 December 2011.
is on the pattern and anomalies associated with the storm. The value of EFS and anomalies with EFS data are presented. Ensemble data shown here are from the NCEP Global Ensemble forecast system which is run at 75km in horizontal resolution. The emphasis here is on products which may aid in predicting high wind events. This includes the probability strong winds at various levels to include 10m, 850 hpa, 700 hpa and 500 hpa. These data were also used to examine the pattern using the 27.5km NCEP GFS 00-hour forecasts. The pattern and standardized anomalies followed the methods outlined in Hart and Grumm (2001) and the GFS 00-hour forecasts were used to establish the pattern and standardized anomalies. The term R-Climate is used in reference to analysis and forecast which use re-analysis climate data to diagnose or forecast the departures from normal. Several European Center for Medium Range Forecasting (ECMWF) products were used to show the potential for high over the British Isles. Products from the ECMWF 50 member ensembles EFI are shown. Data include both plan view and point data. The EFI was developed as described by Lalaurette (2003). Forecasts from each member are compared to the internal ensemble climatology. According to the ECMWF website, an EFI approaching one would imply that most members a predicting values higher than those in the 18 year climatology. M-climate, the model internal climatology is used to compare the current forecasts to the model climate. This allows the forecaster to understand and see when the model is forecasting a high end event. In this study, the focus is on EFI products related to high winds. At the current time, no rainfall data for New Zealand is available. Observed amounts herein are currently based solely on web-based news accounts. For brevity times are presented as day and hour in the format 13/1200 UTC and 14/0000 UTC which would be 1200 UTC 13 December and 0600 UTC 14 December 2011 respectively. Fully qualified dates are limited to comparative data from times outside of 1 and 31 December 2011. 3. The Storm system and impacts i. The large scale pattern The circulation about the strong ridge northeast of New Zealand and the strong trough approaching New Zealand (Fig. 1) played a critical role in transporting tropical moisture, as indicated by the precipitable water (Fig. 2) into the South Island. The deep 500 hpa trough (Fig. 1) was associated with a deep surface cyclone over the Tasman Sea (Fig. 4a-e) and a strong surface anticyclone over the southern Pacific (Fig. 4a-e). Despite the fact that the cyclone would weaken as it approached New Zealand, the isobars implied strong northerly flow between these to pressure systems. ii. Regional pattern and key anomalies The regional perspective shows that the surface cyclone central pressure was about -3s below normal and the cyclone deepened to about 987 hpa by 13/1800 UTC (Fig. 5b) and it moved southward toward western New Zealand (Fig. 5a-f). The strong northerly flow (Fig. 6) brought 850 hpa winds of 40 to 50 kts which were in the 3 to 5 above normal range. The GFS did show some 5-6 above normal winds at 14/0000 UTC.
The 850 hpa v-winds (Fig. 7) showed a persistent period of 4 to 6 above normal northerly winds between 14/0000 and 14/1200 UTC (Figs. 7c-e). The strong low-level northerly flow brought a plume of deep moisture into New Zealand (Fig. 8). The PW values were on the order of 40 to 45 mm which produced mainly 3-4 above normal PW anomalies. However, near the coast of the South Island the PW anomalies were on the order of 4-5 above normal from 13/1800-14/0000 UTC (Fig. 8b-c). The plume of deep moisture and strong northerly flow produced high values of moisture flux (MFLUX) at 850 hpa (Fig. 9). MFLUX anomalies of over 6 were observed along the northern end of the South Island, where the heavier rain was reported in news reports, at 14/0000 UTC (Fig. 9c). The high MFLUX values, in the 3 to 6 range persisted over this region from about 14/0000 through 14/1200 UTC. These data imply the presence of a well-known heavy rainfall pattern over New Zealand, focused on the South Island. iii. Forecasts The GFS forecasts for the total QPF for the period ending at 14/1200 UTC (Fig. 10) and a 12km and 4km WRF initialized with the GFS (Fig. 11) shows the general areas of heavier precipitation in New Zealand. These forecasts show that rainfall in excess of 175 mm was predicted by all the runs in northern portions of the South Island. The 27.5km GFS run from 13/0600 (Fig. 10b) predicted over 200 mm of QPF in the region were the heaviest rain and flooding was reported. The higher resolution WRF runs (Fig. 11) show that the 12km WRF predicted over 300 mm of QPF (Fig. 11a) in the general region of New Zealand where the heaviest rainfall was observed. The forecasts also showed heavy rainfall, over 200 mm along the ridge axis farther south along the north-coast of the South Island. The 4km WRF (Fig. 11b) showed lower amounts and more focused amounts in similar areas relative to the 12km run. It is unclear as to why the 4km run produced lower QPF amounts relative to the 12km run. GEFS forecasts of the probability of 75 mm or more QPF from forecasts initialized at 12/1200, 12/1800 and 13/0000 UTC (Fig. 12) show some of the limits of coarser models and ensemble forecast systems when predicting heavy rainfall. These data also likely show the impact of QPF spin-up in models and coarser global models in general. The 13/0000 UTC forecast has lower probabilities and lower amounts relative to the 12/1200 and 12/1800 UTC cycles. Overall, the GEFS had the concept of heavy rainfall well predicted. Various cycles were examined and an effort was made to show the 100 mm predictions. However, the 75 mm values were more illustrative and shown here. iv. Extreme forecast Indices NCEP GEFS R-Climate based anomalies and mean fields from 10/1200 and 12/1200 UTC (Figs. 13 & 14) show the pattern for heavy rainfall with a high probability of anomalies being in excess of 2s above normal. Both cycles show a high probability of a heavy rainfall pattern over the eastern end of the South Island.
The EC EFI (Fig. 15) using forecast QPF from each member and the internal QPF M-Climate showed a high probability of an internal model based climate heavy rainfall event. Clearly the ensemble system was predicting a high end QPF event, which translated to a high end QPF and flood event in the real atmosphere. v. Observations Rainfall is shown in Figure 16. These data were obtained from news reports and are focused over the region of heaviest rainfall. This is also the region where most of the flood, mudslide, and water damage was observed. 4. Conclusions A high impact heavy rainfall event impacted the South Island of New Zealand on 13-14 December 2011. News reports suggested that the worst flooding and heaviest rainfall fell in the northern reaches of the South Island. V-winds with high PW plumes are a typical heavy rainfall event type. In the southern hemisphere, this requires anomalous northerly winds as opposed to the strong southerly winds often observed in the northern hemisphere. The key point is the anomalous poleward flow of deep moist air into the region affected by the rainfall. The deep plume of moisture, with over 3 PW anomalies, and the high moisture flux values are features often associated with many of the high end or extreme rainfall events. The climate data here help reinforce that this was a well-established pattern favoring rainfall and the R-Climate anomalies aided in identifying this as a meteorologically and climatologically significant event. Not surprisingly, this was likely one of the most significant rainfall events in New Zealand in decades. The model QPFs suggest that higher resolution often pays dividends at shorter forecast ranges. The NCEP GEFS at 75km had difficulty producing probabilities greater than 60% for 100 mm of QPF and thus 75 mm predictions where shown (Fig. 12). The finer resolution GFS had more than double the QPF in most of the GEFS members (Fig. 10). The 12km WRF, initialized using the GFS, produced over 300 mm of precipitation in northernmost areas of the South Island, in close proximity to where the heaviest rainfall (500 mm) and worse flooding was reported. The 4km WRF did not produce as much QPF as the 12km version. The NCEP GEFS forecasts appeared to show higher QPF amounts when initialized 12 or more hours prior to the period of heavy rainfall. This may reflect a precipitation spin-up issue within the coarser modeling system. This affect was not as clear in 13/0000 and 13/0600 UTC forecasts shown. 5. Acknowledgements The EC EFI was provided by the ECMWF and specifically sent by Thomas Petroliagas. Thomas provided many images, 3 of which were used in Figure 15. Craig Evanego of the NWS in State College gathered news reports and plotted rainfall amounts used in Figure 16. 6. References
British Broadcasting Corporation 2011Worst floods in New Zealand for 10 years (and similar stories 14-15 December 2011). Carlson, T.N 1980: Airflow through Midlatitude Cyclones and the comma head pattern. Mon. Wea. Rev.,108,1498-1509. (See figure 4). Graham, Randall A., and Richard H. Grumm, 2010: Utilizing Normalized Anomalies to Assess Synoptic-Scale Weather Events in the Western United States. Wea. Forecasting, 25, 428-445. Grumm, R.H 2011: The Central European and Russian Heat Event of July-August 2010.BAMS, 92, 1285-1296. Grumm, R.H. and R. Hart. 2001: Standardized Anomalies Applied to Significant Cold Season Weather Events: Preliminary Findings. Wea. and Fore., 16,736 754. Hart, R. E., and R. H. Grumm, 2001: Using normalized climatological anomalies to rank synoptic scale events objectively. Mon. Wea. Rev., 129, 2426 2442. Junker, N.W, M.J.Brennan, F. Pereira,M.J.Bodner,and R.H. Grumm, 2009:Assessing the Potential for Rare Precipitation Events with Standardized Anomalies and Ensemble Guidance at the Hydrometeorological Prediction Center. Bulletin of the American Meteorological Society,4 Article: pp. 445 453. Junker, N. W., R. H. Grumm, R. Hart, L. F. Bosart, K. M. Bell, and F. J. Pereira, 2008: Use of standardized anomaly fields to anticipate extreme rainfall in the mountains of northern California. Wea. Forecasting,23, 336 356. Lalaurette, F. 2003: Early Detection of abnormal weather conditions using a probabilistic extreme forecast index. QJRMS,129,3037-3057. Neiman, P.J., F.M. Ralph, G.A. Wick, J. D. Lundquist, and M. D. Dettinger, 2008: Meteorological characteristics and overland precipitation impacts of atmospheric rivers affecting the west coast of North America based on eight years of SSMI/satellite observations. J. Hydrometeor., 9, 22-47. Root, B., P. Knight, G.S. Young, S. Greybush, R.H. Grumm, R. Holmes, and J. Ross, 2007: A fingerprinting technique for major weather events. Journal of Applied Meteorology and Climatology, 46, 1053 1066. Stuart N. A., and R. H. Grumm, 2006: Using wind anomalies to forecast east coast winter storms. Wea. and Forecasting, 21, 952-968.
Figure 1. NCEP GFS 00-hour forecasts of 500 hpa heights (m) and 500 hpa height anomalies (s) in 12-hour periods from a) 0000 UTC 13 December 2011 through f) 1200 UTC 15 December 2011. Heights every 60 m. Return to text.
Figure 2. As in Figure 1 except for precipitable water (mm) and precipitable water anomalies. Preciptable water is every 5mm. Return to text.
Figure 3. General Map of the Islands and towns of New Zealand. Map from the following site: http://www.ezilon.com/maps/images/oceania/new-zealand-map.gif. Return to text.
Figure 4. As in Figure 1 except for mean sea level pressure (hpa) and pressure anomalies. Isoabars every 4 hpa. Return to text.
Figure 5. As in Figure 4 except zoomed in over New Zealand and in 6-hour increments from a) 1200 UTC 13 through f) 1800 UTC 14 December 2011. Return to text.
Figure 6. As in Figure 5 except for 850 hpa winds (kts) and total wind anomalies. Return to text.
Figure 7. As in Figure 5 except for 850 hpa winds (kts) and v-wind anomalie.. Return to text.
Figure 8. As in Figure 5 except for precipitable water and precipitable water anomalies.. Return to text.
Figure 9. As in Figure 5 except for 850 hpa moisture flux (g/kg ms-1) and moisture flux anomalies... Return to text.
Figure 10. NCEP GFS forecasts of total accumulated precipitation (mm) for the period ending at 1200 UTC 14 December 2011 showing the forecasts from GFS cycles initilzed at a) 0000 UTC 13 December and b) 0600 UTC 13 December 2011. Shading as in color scale and contours every 25 mm beginning at 200 mm. Return to text.
Figure 11. As in Figure 10 except for a) 12km WRF and b) 4km WRF forecasts of total QPF initialized at 0600 UTC 13 December 2011 from the 0600 UTC NCEP GFS as boundary and lateral boundary conditions. Contours show 200 and 300 mm. Return to text.
Figure 12. NCEP 75km GEFS forecasts of QPF from forecasts initialized at a & d) 1200 UTC 12 December 2011, b &d) 1800 UTC 12 December 2011, and c &f) 0000 UTC 13 December 2011. Upper panels show each cycles mean 75 mm contour and the probability of exceeding 75 mm or more of QPF. The lower panels show the ensemble mean QPF and each member 75 mm contour. Return to text.
Figure 13. NCEP GEFS forecasts initialized at 0000 UTC 10 December 2011 showing a) the mean 850 hpa winds and the probability of 850 hpa winds 2.5 above normal, b) 850 hpa V-winds and the probability of V-winds over 3 s above normal, c) the mean precipitable water and the probability of the precipitable water being over 3 above normal, and d) mean sea level pressure and the probability of the pressure being -2.0 below normal. Return to text.
Figure 14. As in Figure 13 except forecasts initialized at 1200 UTC 12 December 2011. Return to text.
Figure 15. European Center EFI for heavy rainfall from ensembles forecasts initialized (top) 1200 UTC 10 December, (middle) 1200 UTC 11 December and (lower) 1200 UTC 12 December 2011. Return to text.
Figure 16. Rainfall totals based on news reports and for sites for which latitude and longitude could be recovered using google.map. The region is in the northeastern portions of the South Island where the heaviest rain and most damage was observed. Return to text.