anthony kiem and his colleagues - bureau of meteorology · anthony s. kiem,1 callum twomey,1...

Kiem. Journal of Southern Hemisphere Earth Systems Science (2016) 66:162–176

Corresponding author: Anthony Kiem, Centre for Water, Climate and Land (CWCL), Faculty of Science & Information Technology, University of Newcastle, Callaghan, New South Wales, Australia Email: [email protected]

Links between East Coast Lows and the spatial and temporal variability of rainfall

along the eastern seaboard of Australia

Anthony S. Kiem,1 Callum Twomey,1 Natalie Lockart,2 Garry Willgoose,2 George Kuczera,2 AFM Kamal Chowdhury,2 Nadeeka Parana Manage2 and

Lanying Zhang2

1 Centre for Water, Climate and Land (CWCL), Faculty of Science & Information Technology, University of Newcastle, Callaghan, New South Wales, Australia

2 School of Engineering, University of Newcastle, Callaghan, New South Wales, Australia

(Manuscript received November 2015; accepted June 2016)

East Coast Lows (ECLs) are intense low-pressure systems which occur over the subtropical east coasts of southern and northern hemisphere continents. ECLs are typically associated with gale force winds, large seas, storm surges, heavy rainfall and flooding. While ECL impacts are typically seen as negative the rainfall asso-ciated with ECLs is also very important for urban water security within the heavi-ly populated eastern seaboard of Australia (ESA). This study investigates histori-cal ECLs to gain insights into the timing, frequency, intensity and location of ECL occurrence as well as the magnitude and spatial extent of ECL impacts on rainfall. The different characteristics and impacts associated with different ECL sub-types are highlighted and it is proposed that this spatial and temporal variability in ECL behaviour at least partially explains why the ESA is hydroclimatically different to the rest of Australia and why different locations within the ESA have such differ-ent rainfall patterns—and therefore different levels of flood and drought risk. The-se insights are critical to the objectives of the New South Wales government fund-ed Eastern Seaboard Climate Change Initiative (ESCCI), in particular Project 5 which focuses on the water security impacts of ECLs. The results of this work will be used to produce climate-informed stochastic daily rainfall simulations that are more realistic than existing stochastic rainfall simulation methods at preserving the statistics important for catchment-scale hydrology (e.g. clustering of extreme events, long-term persistence, frequency/duration/magnitude of wet and dry spells). These simulated rainfall sequences, that incorporate the spatial and tem-poral hydroclimatic variability caused by ECLs and other climate phenomena, are important inputs into the hydrological models used to determine current and future urban water security within the ESA.

1. Introduction

East Coast Lows (ECLs) are intense low-pressure systems which occur several times a year over the subtropical east coasts of both southern and northern hemisphere continents (e.g. Verdon-Kidd et al. 2010, 2016, Black and Lane 2015, Colle et al. 2015). ECLs often intensify rapidly overnight, making them one of the more dangerous weather systems to affect densely populated coastal regions such as the eastern seaboard of Australia (ESA). ECLs can generate heavy rainfall, often resulting in major flooding, and also gale force winds, resulting in large swells, rough seas, and wind damage to infrastruc-

Kiem. Links between ECLs and eastern seaboard rainfall variability. 163

ture. In fact, Hopkins and Holland (1997) estimated that ECLs are responsible for approximately 7% of major Australian disasters and more recent work by Callaghan and Power (2014) found that at least 57% of all major floods to occur in coastal catchments along the ESA from 1860–2012 were associated with ECLs.

In addition to their negative impacts, ECLs and their associated rainfall have also been identified as important for generat-ing significant inflows into major water storages along the east coast of Australia (e.g. Pepler and Rakich 2010, Verdon-Kidd et al. 2010, 2016, Kiem and Twomey 2014, Pepler et al. 2014a, 2014b, Black and Lane 2015). Each year there are about between ten and thirty significant maritime lows and generally (i.e. at least a few times a year), these troughs devel-op into a major ECL event. These major ECL events are critical to water security within the ESA, as they refill reservoirs. Unfortunately, ECLs have high spatial and interannual variability, with some years and locations experiencing ten or more ECLs while in other years and locations only a few will develop. Little is known as to what causes this variability, the range of variability experienced in the past, and/or how ECL behaviour might change in the future.

This region where ECLs impacts are typically experienced contains about half of Australia’s population, and a proportionate amount of economic infrastructure and activity. Yet despite the apparent importance of ECLs, understanding into the frequen-cy, intensity, duration, and location of their impacts (and how and why that varies over space and time) and how to deal with them is limited. To further complicate matters, emerging research reveals that there are several different ECL sub-types, each forming differently and with distinctive characteristics (e.g. Speer et al. 2009, Browning and Goodwin 2013). Good examples of how little we know about ECLs and how vulnerable we are to their impacts include the significant impacts (including loss of lives) associated with the ECL related storms and flooding that occurred in April 2015 and also the ECL events that oc-curred in June 2007 and resulted in extensive flooding, storm damage and erosion, and also the grounding of the Pasha Bulker (a 40 000 tonne bulk carrier ship) on Nobbys Beach, near Newcastle harbour. At the time the ‘Pasha Bulker storm’ was the 4th largest insurance loss in Australia since 1968 and this generated interest in ECLs at all levels of government, most notably in terms of their potential impact and future variability in their behaviour under anthropogenic climate change (e.g. Verdon-Kidd et al. 2010, 2016). This recognition across government largely motivated the development of the Eastern Seaboard Cli-mate Change Initiative (ESCCI; http://climatechange.environment.nsw.gov.au/Impacts-of-climate-change/East-Coast-Lows/Eastern-Seaboard-Climate-Change-Initiative).

The New South Wales (NSW) government funded ESCCI commenced in 2007 to address the knowledge gaps associated with the causes and impacts of ECLs and to better understand current and future ECL related risks. ESCCI aims to under-stand the climate of the ESA (roughly the area between the east-coast and the Great Dividing Range; see

Figure 1), the effect of climate change on the ESA, and the implications for climate change adaptation in the ESA. A major motivation for ESCCI is that the relatively narrow (~100 km east to west) coastal strip is poorly resolved in current genera-tion global climate models (grid resolution of ~250 km x 250 km) so it has proven difficult to infer climate change trends and impacts for the region. Yet we know from studies of natural climate variability that the hydroclimate east of the Great Dividing Range responds differently to large-scale ocean drivers than regions west of the Divide (e.g. Verdon and Franks 2005, Verdon-Kidd and Kiem 2009, Timbal 2010). Therefore, a strong case can be made that climate change will influ-ence the ESA in unique ways, requiring focused study in its own right.


Figure 1 The eastern seaboard of Australia (ESA), shown on the left, is roughly the area between the east coast and the Great Dividing Range.


The first program within ESCCI, the ESCCI-ECL research program, commenced in 2010 and consists of a suite of inter-related projects led by the NSW Office of Environment and Heritage. The ESCCI-ECL research program aims to improve our understanding of past, current and future ECLs and assess how these events influence extreme rainfall events, coastal processes and water security. The five projects within ESCCI-ECL arose from extensive workshopping since 2008 with researchers and stakeholders at all levels of government and industry and are listed and described here: http://climatechange.environment.nsw.gov.au/Impacts-of-climate-change/East-Coast-Lows. This paper focuses on research conducted as part of ESCCI-ECL Project 5 (i.e. regional water security impacts of extreme ECLs on ESA reservoirs) which aims to better understand the following knowledge gaps:

• The relationship between ECLs and the intensity, frequency, duration and location of extreme historical storm events (i.e. heavy rainfall, high seas and/or strong winds).

• How ECLs form (i.e. what makes a regular maritime low turn into a major ECL event) and the characteristics and relative importance of different types of ECLs (i.e. ECL ‘bombs’ versus others).

• Historical (i.e. instrumental and pre-instrumental/paleo) ECL intensity, frequency, location, duration, path and geographical extent and how this and the related impacts have varied (and/or trended) over time.

• Historical ECL intensity, frequency, location, duration, path and geographical extent and how this and the related impacts are linked to large-scale ocean-atmospheric processes such as El Niño/Southern Oscillation (ENSO), In-terdecadal Pacific Oscillation (IPO), Indian Ocean Dipole (IOD), Southern Annular Mode (SAM) and subtropical ridge (STR) (refer to Risbey et al. 2009, Gallant et al. 2012, and Kiem and Verdon-Kidd 2013 for further details on these large-scale ocean-atmospheric processes and what we know about their relationship with hydroclimatic variability in Australia).

• Historical variability and trends in ECL behaviour and how this has impacted historical water availability, dry spells, and breaking of dry spells within the ESA.

• The likelihood and characteristics of changes to ECL intensity, frequency, location, path and geographical extent under anthropogenic climate change and resulting implications for water security within the ESA.

In this paper we focus specifically on the following questions:

• When do ECLs occur and has this changed over time? • Where do ECLs occur and has this changed over time? • What is the relationship between ECLs and rainfall at different locations within the ESA?

2. Data and method

2.1 Defining and categorising ECLs

In this paper we use the East Coast Cyclones (ECC) dataset produced by Browning and Goodwin (2013) to obtain a time series of ECC (referred to as ECLs in this paper) occurrences and locations from 1979 to 2011. The Browning and Good-win (2013) ECC dataset objectively identifies, tracks and classifies ECLs from the 1.5° European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis Interim (ERA-Interim) 6-hourly sea level pressure (SLP) data. Con-sistent with the Australian Bureau of Meteorology (BOM) definition of an ECL, the Browning and Goodwin (2013) algo-rithm initially identifies all closed contour low pressure systems in the study area by measuring pressure gradients. This initial ECL identification is then refined to eliminate weaker or less significant closed circulation lows and by treating two or more closed circulation lows that occur very close to each other in time (sequential 6-hourly time steps) or space (3–6° latitude-longitude) as a single ECL event. The Browning and Goodwin (2013) method identifies significantly more events than have been identified in previous subjective analyses (e.g. Speer et al. 2009) but, importantly, Ji et al. (2015) show that the major ECL events are consistent when comparisons are made between the Browning and Goodwin (2013) and Speer et al. (2009) ECL datasets.

Browning and Goodwin (2013) then look at ‘pre-storm trajectories’ to objectively sub-classify each ECL event into one of five main ECL sub-types previously defined in Public Works Department (1985, 1986). This sub-classification is based on: storm-track latitude, whether the storm evolved mostly over the land or the sea, whether it tracked north or south, and its orientation to Australia’s Great Dividing Range). Accordingly, the ECL sub-types are defined as follows:

1) Easterly trough lows (ETL): events that track mostly east of the Great Dividing Range and in a southerly direction.


2) Southern secondary lows (SSL): events that track mostly over the ocean and in a northerly direction. 3) Inland troughs (IT): events that evolve mostly over land, west of the Great Dividing Range and north of 30°S. 4) Continental lows (CL): events that evolve mostly over land, west of the Great Dividing Range and south of 30°S. 5) Extratropical cyclones (XTC): differentiation between ETLs and ECLs that evolve from tropical cyclones based

on storm track trajectories is problematic, as both storm types develop in a similar region and follow similar tracks. However, the BOM maintains a database of all tropical cyclone occurrences and tracks from 1900 to pre-sent (http://www.bom.gov.au/cyclone/history/). XTCs are therefore defined as events that evolved from storm systems included in the BOM tropical cyclone database.

The spatial domain for our study, and the method for defining and categorising ECLs, is the same as that used by Brown-ing and Goodwin (2013)—resulting in the same ECL event characteristics (i.e. intensity, frequency, location, duration, path, geographical extent). This is deliberate, and necessary, as part of the reason why so much uncertainty and so many knowledge gaps remain around ECLs and their impacts is because of the multiple ECL definitions and datasets that exist (there is inconsistency and contradiction in the published literature on ECLs and their impacts and this is at least partially related to the fact that it is rare to find two studies that use the same ECL definitions or dataset; e.g. Speer et al. 2009, Pep-ler and Rakich 2010, Browning and Goodwin 2013, Callaghan and Power 2014, Pepler et al. 2014a, 2014b, Black and Lane 2015). One of the first priorities of ESCCI-ECL was to agree on an ECL definition and dataset that all ESCCI-ECL projects should use and it was determined that the Browning and Goodwin (2013) was the most appropriate. However, even though our study is based on the ECL dataset prepared by Browning and Goodwin (2013) the objectives and method-ology of our study differs in the following ways:

• Browning and Goodwin (2013) focus on winter (May to August) but we investigate ECL behaviour in all months. • Browning and Goodwin (2013) focus on the number of storm days but we investigate frequency of ECL events. • Browning and Goodwin (2013) focus on mean storm tracks for ECLs from 48 hours before to 48 hours after the

storm peak. We also do this but expand by also showing information about the distribution of storm locations within each ECL sub-type (i.e. not just the mean storm track) in the various stages of the ECL development and decline. This information is important for hydrology as we want to know when (i.e. pre- or post-peak intensity and peak impact) and which ECLs spend most time over the water supply catchments within the ESA.

• Browning and Goodwin (2013) do not investigate the relationship between ECLs and rainfall within the ESA.

2.2 Relationship between ECLs and ESA rainfall

To assess the relationship between ECLs and ESA rainfall, observed rainfall data was obtained from BOM high quality gauges located within the ESA for the period where the Browning and Goodwin (2013) ECL sub-type information was available (1979–2011). To allow easier visualisation of spatial patterns, gridded daily rainfall data from the Australian Water Availability Project (AWAP; Jones et al. 2009), a joint initiative of the BOM and the Commonwealth Scientific and Industrial Research Organisation (CSIRO) was also used (refer to Tozer et al. 2012 for discussion on why both observed rainfall data, i.e. obtained directly from rainfall stations, and gridded AWAP data should be used).

The daily rainfall data was first stratified based on whether or not the day was associated with an ECL event (occurring anywhere from ‘0’ hour on within the spatial domain used by Browning and Goodwin (2013)) and then the ECL rainfall days were further stratified according to the sub-type the ECL was classified as. Two limitations of this method are: (1) no consideration was given to the location of the ECL within the study domain (i.e. rainfall in the ESA is considered ECL-related if an ECL is occurring anywhere in the study area) and (2) if two or more ECLs are occurring on a given day a de-cision has to be made as to which ECL the rainfall for that day is allocated to. The first limitation we have not addressed in this study so as to keep things as simple as possible (i.e. no consideration is given to the distance between rainfall location and the location of an ECL occurring at the same time) for the stochastic rainfall generation, done in parallel studies, that these results are used for (subsequent studies will assess the sensitivities and uncertainties associated with this decision). Regarding the second limitation, over the entire 1979–2011 study period only 20 days were associated with two ECL events (and no days were associated with more than two ECL events) out of a total 1290 days associated with an ECL (corresponding to only 1.55% of all ECL days). When two ECLs did occur on the same day it was usually when the peak or end of a ‘primary’ ECL event overlapped with the beginning of a ‘secondary’ ECL event. Furthermore, of the 20 days that contained two separate ECLs, 12 of these days contained two ECLs of the same sub-type. For the residual 8 days (0.62% of all ECL days) it was assumed that rainfall occurring over the ESA would most likely be associated with the


‘primary’ event (noting again that this only occurred for 0.62% of all ECL days so is unlikely to have a significant effect on the results or conclusions).

In this paper, to aid synthesis of results we only present and discuss the results obtained using the Browning and Goodwin (2013) ECL sub-types and AWAP rainfall data (see Section 3.3) but readers are referred to Twomey and Kiem (2015, 2016a) for more detailed analyses obtained using multiple ECL datasets and both gauged and gridded rainfall. Importantly, while differences emerge as a result of using different ECL or rainfall datasets the conclusions and key insights made in this paper do not change. However, this paper focuses only on rainfall patterns and large-scale changes in ECL behaviour so care should be taken, for example, when using the insights here to develop inputs for hydrological modelling as the AWAP dataset has documented limitations and uncertainties, especially at higher elevations and/or in areas or time periods where gauged information is sparse (e.g. Jones et al. 2009, Tozer et al. 2012, King et al. 2013).

3. Results

3.1 When do ECLs occur and has this changed over time?

Figure 2 shows the frequency of occurrence for each of the five different ECL sub-types within the different months of the year. Overall, Figure 2 shows that ECLs are possible in any month of the year but are most frequent from April to October. However, while it is true that ECLs can occur in any month of the year it is important to note that some ECL sub-types are far more prevalent at certain times of year. ETL, SSL, and CL sub-types are the most common but this varies markedly across the different months (e.g. the frequency of ETLs is lower in the second half of the year (Jul–Dec), the frequency of CLs is lower in the first half of the year (Jan–Jun), and SSLs rarely occur in summer (Dec–Jan–Feb)). Following on from this, while each ECL sub-type is more likely to occur in certain months,

Figure 2 also shows that, except for XTCs (which have only occurred from November to April), any ECL sub-type can occur at any time of the year. This is an important observation since it is the ECL events that occur ‘out of season’ that potentially cause the most problems for water resources management due to their rare and unexpected impacts for that time of year (e.g. reservoir filling events occurring at a time of the year when reservoir levels are typically already high).

Figure 2 Monthly frequency of ECLs, and ECL sub-types, from 1979–2011. ECL classification and timing is as per the Browning and Goodwin (2013) ECL dataset which covers the period 1979–2011.


Figure 3 shows the frequency of ECL events, and also the breakdown into ECL sub-types, over the period 1979–2011. Figure 3 clearly shows that there is large interannual variability in the number of ECLs that occur each year and that this applies to ECLs over all and the different ECL sub-types. Relationships between this interannual variability in the fre-quency of ECL events and large-scale climate processes (e.g. ENSO etc.) have been suggested and this is the topic of re-cent and ongoing research (e.g. Gallant et al. 2012, Pepler et al. 2014a, 2014b, Twomey and Kiem 2015, 2016a, Power and Callaghan 2016, Tozer et al. 2016).

Figure 3 Annual frequency of ECLs, and ECL sub-types, from 1979–2011.

Figure 4 shows the results of investigations into whether or not any trends exist in the frequency of ECL occurrence (ECLs overall and the different ECL sub-types). It should first be noted that 1979–2011 (i.e. the period covered by the Browning and Goodwin (2013) ECL dataset used here) is most likely not a long enough time period to detect anything meaningful in terms of trends (especially given the large interannual variability in ECL occurrence that is illustrated in Figure 3). Never-theless, since there have been suggestions made in the mainstream media that the frequency of storms has been increasing we investigated the data to see if there was any evidence to support these claims.

Figure 4 shows no obvious or statistically significant trend in either the total number of ECLs that have occurred per year or the number of occurrences per year for each of the ECL sub-types. This is consistent with Ji et al. (2015) and Twomey and Kiem (2016a) who performed a similar analyses using other ECL datasets (e.g. Speer et al. 2009 and the BOM’s Maps and Tables of Climate Hazards on the Eastern Seaboard; MATCHES) and also concluded no statistically significant trends in annual ECL frequency. It is important to note that, as previously mentioned, the periods covered by the available ECL sub-type datasets (36 years for Speer et al. 2009; 32 years for Browning and Goodwin 2013) are short in the context of detecting and quantifying meaningful trends. Therefore, since the observations are limited the results in Figure 4 do not necessarily mean that no trend exists over the longer term (e.g. from 1900–present) and also do not exclude the possibility that the annual frequency of ECLs or the ECL sub-types will change in the future. However, based on the evidence cur-rently available it appears that annual ECL and ECL sub-type frequency follows more of a cyclical pattern which provides more weight to the suggestion that ECL behaviour (i.e. frequency, timing) is related to large-scale processes known to be associated with interannual to multidecadal hydroclimatic variability in Australia (e.g. Gallant et al. 2012, Pepler et al. 2014a, 2014b, Twomey and Kiem 2015, 2016a, Power and Callaghan 2016, Tozer et al. 2016).


Figure 4 Trend analysis of ECLs, and the ECL sub-types, from 1979–2011.

3.2. Where do ECLs occur and has this changed over time?

Figure 5 shows the spatial variation in the location of ECLs, categorised by ECL sub-type as per Section 2.1, that have occurred from 1979–2011 during four different phases of the ECL event. Also included in Figure 5 is the spatial variation in the location of minimum mean sea level pressure (MSLP) for each ECL sub-type, which is indicative of the location of peak intensity for each ECL event. Note that the minimum MSLPs (i.e. peak intensity) for each ECL sub-type are more tightly clustered than the location of the ECLs during the four different phases of the ECL event. This means it is possible to present the map of minimum MSLPs for each ECL at a different scale (zoomed in) to the other four maps, making it easier to visually distinguish the differences between different ECL sub-types.

As with the previous results relating to the frequency and timing of ECLs it is clear that the different ECL sub-types also have different characteristics when it comes to the location that they typically occur in. Some geographical spread in the locations of different ECL sub-types is expected due to the way the ECL sub-types are defined (see Section 2.1) however the spatial variability illustrated in Figure 5, far more than what is suggested by the mean storm track shown in Figure 1 of Browning and Goodwin (2013), is particularly important for water resource management as even a small shift in the loca-tion of ECL events could mean that reservoir filling rainfall no longer falls over the catchment areas for important urban water supply systems within the ESA. Also important is the fact that some regions within the ESA are clearly influenced by multiple ECL sub-types (e.g. the regions under overlapping boxes in the entire ECL life-cycle, i.e. (iv), part of Figure 5). Therefore, the results in Figure 5 demonstrate that it is crucial for water resources management and infrastructure de-sign to know which ECL sub-types are important for which regions within the ESA and how the location of those im-portant ECL sub-types has changed over time (and may change in the future).

Figure 5 also reveals that there is not much difference between categories (ii) and (iii), indicating that not many ECLs persist for longer than 48 hours after their detection (i.e. day 0). This is a result that was expected but was necessary to confirm as it is im-portant to know when (i.e. pre or post peak intensity and peak impact) and which ECLs spend most time over the water supply catchments within the ESA—these results demonstrate that it is likely that ECLs have minimal impact 48 hours after their devel-opment and this is useful information when forecasting ECL impacts or stochastically simulating rainfall for use in water re-source management, assessing water security, and/or infrastructure design. The results in Figure 5 also serve as useful baseline information with which to compare past or future ECL behaviour (either modelled or observed) to determine whether the loca-tion and/or persistence of ECLs or their impacts is different to what has occurred in the 1979–2011 period studied here.


Figure 5 Spatial variation in the location of ECLs that have occurred from 1979–2011 during four different phases of the ECL event: (i) 48 hours prior to development of the ECL event (top left), (ii) 48 hours after development of the ECL event (top right), (iii) from development of the ECL event until ECL no longer exists (middle left) and (iv) over the entire event life-cycle, i.e. (i) to (iii) combined (middle right). Also included is the spa-tial variation in location of minimum mean sea level pressure (MSLP) for each ECL sub-type, i.e. location of peak intensity (bottom left). Note that the minimum MSLPs for each ECL sub-type map uses a different scale to the other four maps.


3.3 What is the relationship between ECLs and rainfall within the ESA?

Figure 6 shows the average, standard deviation and range of daily rainfall associated with the five ECL sub-types at differ-ent locations along the ESA. The intention here is to illustrate the distribution of rainfall associated with each ECL sub-type across the different parts of the ESA. In addition, to illustrate the relative importance of different ECL sub-types for different parts of the ESA, Figure 7 shows the rainfall associated with each ECL event, stratified by ECL sub-type, as a percentage of total ECL rainfall and as a percentage of all rainfall (i.e. ECL and non-ECL related rainfall) at different loca-tions within the ESA.

Generally, Figure 6 and Figure 7 show that ETLs are associated with the largest amount of rainfall and impact most re-gions of the ESA, except the south ESA. SSLs and CLs are responsible for a lot of rainfall in the south ESA but have less influence elsewhere. ITs affect the whole ESA but their overall impact in terms of rainfall magnitude is not as great as the other ECL sub-types. XTCs most influence the mid-north ESA and are associated with significant rainfall when they occur but given there are not as many XTCs as there are other ECL sub-types the amount of rainfall associated with XTCs is relatively small when compared to total ECL and non-ECL rainfall received. As expected, the spatial patterns of the ECL sub-type and ESA rainfall relationships is strongly linked to the spatial variability in the locations of the different ECL sub-types (Figure 5).

From Figure 6 and Figure 7, it is clear that for a given location within the ESA there is marked variability in the rainfall associated with different ECL sub-types (Figure 6) and also large differences in the proportion of total rainfall ECLs are responsible for (Figure 7). This implies that different ECL sub-types are more/less important for water security at different locations and that it is not enough to focus on how ECLs in general impact ESA’s hydroclimate or might change in the future. Rather, there is a need to identify, on a location by location basis, the ECL sub-type(s) that are most important for a given region followed by a focus on how the characteristics, and impacts, of the critical ECL sub-type(s) have changed in the past and are projected to change in the future. Also important is to understand and quantify the mechanisms responsi-ble for non-ECL related rainfall (since, as shown in Figure 7, while obviously important ECLs are still only responsible for up to ~50% of rainfall in most ESA locations). This further research is currently underway.

The spatial pattern of the different rainfall statistics associated with each ECL sub-type (as illustrated in Figure 6 and Fig-ure 7) also suggest significant inhomogeneity in rainfall patterns across the ESA. The ESA has previously been identified as being different to the rest of Australia in terms of rainfall patterns (e.g. McBride and Nicholls 1983, Drowsdowsky 1993, Verdon and Franks 2005, Verdon-Kidd and Kiem 2009, Timbal 2010, Pepler et al. 2014a, 2014b), but inhomogenei-ty in rainfall patterns within the ESA is not well understood. Figure 6 and Figure 7 show that inhomogeneity in rainfall patterns within the ESA does exist and that, combined with other results presented in this paper, suggests that this inhomo-geneity is related to ECL behaviour. This finding is further explored in Twomey and Kiem (2016b).


Figure 6 Average, standard deviation and range of daily rainfall associated with the five ECL sub-types at different

locations within the ESA (period of analysis is 1979–2011).


Figure 7 Rainfall associated with each ECL event, stratified by ECL sub-type, as (on the left) a percentage of total

ECL rainfall and (on the right) as a percentage of all rainfall (i.e. ECL and non-ECL related rainfall) at dif-ferent locations within the ESA (period of analysis is 1979–2011).


4. Conclusions

In this paper we investigate the spatial and temporal variability of ECLs, specifically the five ECL sub-types defined and categorised by Browning and Goodwin (2013), and the relationship between ECLs and rainfall at different locations within the ESA.

We find that:

1) No trend is evident in the timing or frequency of ECL occurrence (based on analysis of 1979–2011) however there is a large degree of interannual variability that may be related to atmosphere-ocean coupled processes such as ENSO, IPO, IOD, SAM or STR. This finding is the same whether ECLs are considered all together or sepa-rately by ECL sub-type. Further work is needed to develop longer datasets of ECL occurrence and to better un-derstand and quantify temporal variability (or changes) in the timing or frequency of ECLs and ECL sub-types.

2) There is a large degree of spatial variability in the location of different ECL sub-types during the different stages of the ECL life cycle. Some geographical spread in the locations of different ECL sub-types is expected due to the way the ECL sub-types are defined (see Section 2.1) however the spatial variability in the location of different ECL sub-types is particularly important for water resource management as even a small shift in the location of ECL events could mean that reservoir filling rainfall no longer falls over the catchment areas for important urban water supply systems within the ESA. Also important is the fact that some regions within ESA are clearly influ-enced by multiple ECL sub-types—again, demonstrating that it is crucial for water resources management and in-frastructure design to know which ECL sub-types are important for which regions within the ESA and how the location of those important ECL sub-types has changed over time (and may change in the future).

3) For a given location within the ESA there is marked variability in the rainfall associated with different ECL sub-types and also big differences in the proportion of total rainfall ECLs are responsible for. This implies that differ-ent ECL sub-types are more/less important for water security at different locations and that it is not enough to fo-cus on how ECLs in general impact ESA’s hydroclimate or might change in the future. Rather, for water security in specific regions there is a need to identify, on a location by location basis, (a) which ECL sub-type(s) are most important for a given region and how/when they impact and (b) how the characteristics, and impacts, of the criti-cal ECL sub-type(s) have changed in the past (instrumental and pre-instrumental) and are projected to change in the future.

4) The timing, frequency and location of different ECL sub-types appears to be related to the inhomogeneity ob-served in the rainfall patterns across the ESA. This finding is further explored in Twomey and Kiem (2016b).

The spatial and temporal variability in ECL behaviour demonstrated here at least partially explains why the ESA is hydro-climatically different to the rest of Australia and why different locations within the ESA have such different rainfall pat-terns—and therefore different levels of flood and drought risk depending on the overarching climate phase (i.e. frequent ECLs or not, certain ECL sub-types more prevalent than others). These insights are critical to the objectives of the ESCCI, in particular Project 5 which focuses on the water security impacts of ECLs. The results of this work will now be used to produce climate-informed stochastic daily rainfall simulations (e.g. Kiem and Franks 2004, McMahon et al. 2008, Chow-dhury et al. 2015, Mortazavi-Naeini et al. 2015) that are more realistic than existing stochastic rainfall simulation methods at preserving the statistics important for catchment-scale hydrology (e.g. clustering of extreme events, long-term persis-tence, frequency/duration/magnitude of wet and dry spells). These simulated rainfall sequences, that incorporate the spatial and temporal hydroclimatic variability caused by ECLs and other climate phenomena, are important inputs into the hydro-logical models used to determine current and future urban water security in the ESA (Lockart et al. 2016).

5. Acknowledgements

Funding for this research was provided by Australian Research Council Linkage Grant LP120200494 with further funding and/or in-kind support also provided by the NSW Office of Environment and Heritage, Sydney Catchment Authority, Hunter Water Corporation, NSW Office of Water, and NSW Department of Finance and Services.

ESCCI-ECL was funded through the NSW Environmental Trust, the NSW Department of Finance and Services, Hunter Water Corporation and was supported under the Australian Research Council's Linkage Project and Discovery Project schemes. ESCCI-ECL research partners from University of New South Wales (Jason Evans and team), Macquarie Univer-


sity (Ian Goodwin and Stuart Browning) and the Bureau of Meteorology also contributed data and other resources used in the work presented here.

6. References

Black, M.T. and Lane, T.P. 2015. An improved diagnostic for summertime rainfall along the eastern seaboard of Australia. International Journal of Climatology, 35, 4480–4492, doi:10.1002/joc.4300.

Browning, S.A. and Goodwin, I.D. 2013. Large-scale influences on the evolution of winter subtropical maritime cyclones affecting Australia’s east coast. Monthly Weather Review, 141, 2416–2431, doi: http://dx.doi.org/10.1175/MWR-D-12-00312.1.

Callaghan, J. and Power, S.B. 2014. Major coastal flooding in southeastern Australia 1860–2012, associated deaths and weather systems. Australian Meteorological and Oceanographic Journal, 64, 183–213.

Chowdhury, A.F.M.K., Lockart, N., Willgoose, G.R., Kuczera, G., Kiem, A.S. and Parana Manage, N. 2015. Modelling daily rainfall in East Coast of Australia using a Compound Distribution Markov Chain model. Proc. 36th Hydrology and Water Resources Symposium, Hobart, Australia, 7–10 December 2015.

Colle, B.A., Booth, J.F. and Chang, E.K.M. 2015. A review of historical and future changes of extratropical cyclones and associated impacts along the US east coast. Current Climate Change Reports, 1(3), 125–143.

Drosdowsky, W. 1993. An analysis of Australian seasonal rainfall anomalies: 1950–1987. II: Temporal variability and teleconnection patterns. International Journal of Climatology, 13, 111–149. doi: 10.1002/joc.3370130202.

Gallant, A.J.E., Kiem, A.S., Verdon-Kidd, D.C., Stone, R.C. and Karoly, D.J. 2012. Understanding hydroclimate processes in the Murray-Darling Basin for natural resources management. Hydrology and Earth System Sciences, 16, 2049–2068, doi:10.5194/hess-16-2049-2012.

Hopkins, L.C. and Holland, G.J. 1997. Australian heavy-rain days and associated east coast cyclones: 1958–92. Journal of Climate, 10, 621–635.

Ji, F., Evans, J.P., Argueso, D., Fita, L. and Di Luca, A. 2015. Using large-scale diagnostic quantities to investigate change in East Coast Lows. Climate Dynamics, doi: 10.1007/s00382-015-2481-9.

Jones, D.A., Wang, W. and Fawcett, R. 2009. High-quality spatial climate analyses for Australia. Australian Meteorological and Oceanographic Journal, 58, 233–248.

Kiem, A.S. and Franks, S.W. 2004. Multi-decadal variability of drought risk - Eastern Australia. Hydrological Processes, 18(11), 2039–2050.

Kiem, A.S. and Twomey, C.R. 2014. Investigating the relationship between East Coast Lows (ECLs) and east Australian rainfall. Proc. 35th Hydrology and Water Resources Symposium, Perth, Australia, 24–27 February 2014.

Kiem, A.S. and Verdon-Kidd, D.C. 2013. The importance of understanding drivers of hydroclimatic variability for robust flood risk planning in the coastal zone. Australian Journal of Water Resources, 17(2), 126–134, doi:10.7158/W13-015.2013.17.2.

King, A.D., Alexander, L.V. and Donat, M.G. 2013. The efficacy of using gridded data to examine extreme rainfall characteristics: a case study for Australia. International Journal of Climatology, 33, 2376–2387, doi:10.1002/joc.3588.

Lockart, N., Willgoose, G., Kuczera, G., Kiem, A.S., Chowdhury, A.F.M.K., Parana Manage, N., Zhang, L. and Twomey, C. 2016. Case study on the use of dynamically downscaled climate model data for assessing water security in the Lower Hunter region of the eastern seaboard of Australia. Journal of Southern Hemisphere Earth System Science, ESCCI Special Issue, 66, 177–202.

McBride, J.L. and Nicholls, N. 1983. Seasonal relationships between Australian rainfall and the Southern Oscillation. Monthly Weather Review, 111, 1998–2004.

McMahon, T.A., Kiem, A.S., Peel, M.C., Jordan, P.W. and Pegram, G.G.S. 2008. A new approach to stochastically generating six-monthly rainfall sequences based on Empirical Model Decomposition. Journal of Hydrometeorology, 9, 1377–1389.

Mortazavi-Naeini, M., Kuczera, G., Kiem, A.S., Cui, L., Henley, B., Berghout, B. and Turner, E. 2015. Robust optimization to secure urban bulk water supply against extreme drought and uncertain climate change. Environmental Modelling & Software, 69, 437–451, doi:10.1016/j.envsoft.2015.02.021.

Pepler, A. and Rakich, C.S. 2010. Extreme inflow events and synoptic forcing in Sydney catchments. IOP Conference Series: Earth and Environmental Science (EES), 11(012010), doi:10.1088/1755-1315/11/1/012010.

Pepler, A., Coutts-Smith, A. and Timbal, B. 2014a. The role of East Coast Lows on rainfall patterns and inter-annual variability across the East Coast of Australia. International Journal of Climatology, 34, 1011–1021, doi:10.1002/joc.3741.


Pepler, A., Timbal, B., Rakich, C. and Coutts-Smith, A. 2014b. Indian Ocean Dipole overrides ENSO's influence on cool season rainfall across the Eastern Seaboard of Australia. Journal of Climate, 27, 3816–3826, doi:10.1175/JCLI-D-13-00554.1.

Power, S.B. and Callaghan, J. 2016. Variability in severe coastal flooding, associated storms, and death tolls in southeastern Australia since the mid-nineteenth century. Journal of Applied Meteorology and Climatology, 55(5), 1139–1149, doi:10.1175/JAMC-D-15-0146.1.

Public Works Department 1985. Elevated coastal levels: storms affecting NSW coast 1880-1980. Weatherex Meteorological Services and Public Works Department Report 85041, 574 pages, 1985.

Public Works Department 1986. Elevated coastal levels: storms affecting NSW coast 1880-1985. Weatherex Meteorological Services and Public Works Department Report 86026, 5 pages, 1986.

Risbey, J.S., Pook, M.J., McIntosh, P.C., Wheeler, M.C. and Hendon, H.H. 2009. On the remote drivers of rainfall variability in Australia. Monthly Weather Review, 137, 3233–3253, doi:10.1175/2009MWR2861.1.

Speer, M.S., Wiles, P. and Pepler, A.S. 2009. Low pressure systems off the New South Wales coast and associated hazardous weather: Establishment of a database. Australian Meteorological and Oceanographic Journal, 58, 29–39.

Timbal, B. 2010. The climate of the Eastern Seaboard of Australia: a challenging entity now and for future projections. IOP Conference Series: Earth and Environmental Science (EES), 11(012013), doi:10.1088/1755-1315/11/1/012013.

Tozer, C.R., Kiem, A.S. and Verdon-Kidd, D.C. 2012. On the uncertainties associated with using gridded rainfall data as a proxy for observed. Hydrology and Earth System Sciences, 16, 1481–1499, doi:10.5194/hess-16-1481-2012.

Tozer, C.R., Vance, T.R., Roberts, J.L., Kiem, A.S., Curran, M.A.J. and Moy, A.D. 2016. An ice core derived 1013-year catchment-scale annual rainfall reconstruction in subtropical eastern Australia. Hydrology and Earth System Sciences, 20, 1703–1717, doi:10.5194/hess-20-1703-2016.

Twomey, C.R. and Kiem, A.S. (2015): Spatial analysis of Australian seasonal rainfall anomalies and their relation to East Coast Lows on the Eastern Seaboard of Australia. Proc. 36th Hydrology and Water Resources Symposium, Hobart, Australia, 7–10 December 2015.

Twomey, C.R. and Kiem, A.S. 2016a. East Coast Lows (ECLs) and rainfall along the eastern seaboard of Australia (ESA) - comparison of datasets used to record ECL occurrence and impacts on rainfall. Journal of Southern Hemisphere Earth System Science, in prep.

Twomey, C.R. and Kiem, A.S. 2016b. Spatial and temporal variability of Australian rainfall - insights into why the eastern seaboard of Australia is different to the rest of Australia and also internally inhomogeneous. International Journal of Climatology, in review.

Verdon, D.C. and Franks, S.W. 2005. Indian Ocean sea surface temperature variability and winter rainfall: Eastern Australia. Water Resources Research, 41, W09413, doi:10.1029/2004WR003845.

Verdon-Kidd, D.C. and Kiem, A.S. 2009. Nature and causes of protracted droughts in Southeast Australia - Comparison between the Federation, WWII and Big Dry droughts. Geophysical Research Letters, 36, L22707, doi:10.1029/2009GL041067.

Verdon-Kidd, D.C., Kiem, A.S. and Willgoose, G. 2016. East Coast Lows and the Pasha Bulker storm – lessons learned nine years on. Journal of Southern Hemisphere Earth System Science, ESCCI Special Issue, 66, 152–161.

Verdon-Kidd, D.C., Kiem, A.S., Willgoose, G. and Haines, P. 2010. East Coast Lows and the Newcastle/Central Coast Pasha Bulker Storm. Technical report prepared for the National Climate Change Adaptation Research Facility (NCCARF), 68 pages, http://www.nccarf.edu.au/publications/east-coast-lows-and-newcastle-central-coast-pasha-bulker-storm.

anthony kiem and his colleagues - bureau of meteorology · anthony s. kiem,1 callum twomey,1...

Documents