A geomorphic classification of intermittently open/closed estuaries (IOCE) derived from estuaries in Victoria,Australia

Abstract

A large proportion of estuaries along microtidal wave-dominated coastlines worldwide have entrances that intermittently close to the ocean when tidal currents and fluvial discharge are insufficient to erode sediment delivered onshore by waves. In this study, these systems are termed “intermittently open/closed estuaries” (IOCE) in order to include all estuaries which intermittently close to the ocean. IOCE do not fit neatly into existing generalized estuary classification models and have been traditionally recognized as a single estuary type that constitute a rare subset of wave-dominated estuaries. In this study, 111 estuaries in Victoria, Australia, are used to develop a classification model that delineates between different IOCE types. This was undertaken using historic aerial imagery and quantification of the estuary channel width, catchment area, lagoon dimensions and tidal prism derived from remotely sensed data. Field surveying of entrance morphology was undertaken for a subsample of 35 IOCE characteristic of each section of the coast and which had detailed entrance condition records. Using this subset, IOCE were classified into three distinct types using multiple methods of statistical delineation (non-metric multidimensional scaling, hierarchical cluster analysis and distribution analysis). These three types are: (1) Type A, the largest IOCE which both close and open infrequently but for the longest durations; (2) Type B, medium sized IOCE which open and close several times per year for weekly to monthly durations; and (3) Type C (tidal creeks), the smallest IOCE located specifically in high rainfall, mountainous catchments and which exist in a predominantly open state. The three types of IOCE showed an order of magnitude difference in entrance closure duration as controlled by variations in the catchment area, tidal prism volume, dimensions of the estuarine lagoon and the entrance channel at the mouth. The classification is also applicable to wave-dominated coastlines internationally where IOCE are present.

Keywords

Estuary ICOLL classification coastal lagoons TOCE

I Introduction

Estuaries are broadly classified based on the relative dominance of marine and fluvial processes (Dalrymple et al., 1992; Hume and Herdendorf, 1988; Kench, 1999; Kennedy, 2011; Woodroffe et al., 1989). In microtidal regions, wave-dominated estuaries are most common and occur in three forms: (i) drowned river valleys, (ii) barrier estuaries, and (iii) saline coastal lakes (Allen and Posamentier, 1994; Boyd et al., 1992; Roy, 1984). The specific morphology of each type further varies based on the degree of sedimentary infill. Youthful estuaries have limited infill and often deep basins, whereas mature estuaries have little accommodation space in the central basin and tend to be shallow (Heap et al., 2004; Roy, 1984; Saintilan, 2004). The estuary accommodation space refers to the available space for potential sediment accumulation (Slagle et al., 2006). As these systems infill, the entrance often progressively shallows and may close as the accommodation space becomes occupied with sediment. For this reason, entrance condition is often used as a proxy for defining estuarine type (Cooper, 2001; Roy, 1984).

In certain estuarine systems, the entrance is not permanently open to the sea, but oscillates between an open and a closed state (Perissinotto et al., 2010a; Potter et al., 2010; Ranasinghe et al., 1999). Estuaries which intermittently close to the ocean are common on microtidal, wave-dominated coastlines such as southwest and southeast Australia, where they are termed “intermittently closed/open lakes and lagoons” (ICOLLs) (Dye and Barros, 2005; Gale et al., 2006; Haines et al., 2006; Morris and Turner, 2010), South Africa, where they are referred to as “temporarily open/closed estuaries” (TOCE) (Perissinotto et al., 2010b; Whitfield, 1992), New Zealand (Hart, 2009; Kirk and Lauder, 2000), Mexico (Ruiz-Luna and Berlanga-Robles, 2003), south-east Asia (Piazza et al., 2010), and parts of North and South America (Behrens et al., 2013; Hanes et al., 2011; Kraus et al., 2002; Oliveira and Kjerfve, 1993). In this study, we refer to these systems as “intermittently open/closed estuaries” (IOCE), a universally applicable term that covers all estuaries which intermittently close to the ocean. Despite their ubiquity, IOCE are traditionally thought to be a transitional stage of permanently open estuaries on the trajectory to forming a fluvial delta and therefore do not fit neatly into generalized classification models (Cooper, 2001; Roy 1984, Roy et al., 2001).

General estuary classifications based on geomorphic processes are less common than those focusing on the physio-chemical, ecological and management attributes of estuaries. Within existing geomorphic classifications, IOCE are often excluded or recognized merely as a homogenous estuary type that are a rare subset of wave-dominated systems (Boyd et al., 1992; Dalrymple et al., 1992; Hume and Herdendorf, 1988), and there have been few attempts to differentiate between any sub-types of IOCE that might exist. At a continental scale, attempts at differentiating between IOCE sub-types tend to be conceptual classifications only (Cooper, 2001; Jacobs et al., 2011; Kjerfve, 1994). For South African systems, IOCE have been conceptually classified into two sub-types based on the frequency of their connection with the ocean. Intermittently open estuaries (IOE) are closed for > 50% of the time, on average, whereas intermittently closed estuaries (ICE) are closed for < 50% of the time (Bell et al., 2001; Cowley et al., 2001; James et al., 2007; Whitfield and Bate, 2007). In Australia, IOCE are becoming increasingly recognized as an estuary type in their own right (Harris and Heap 2003; Harris et al., 2002), although classification predominantly focuses on estuary ecology, or their degree of modification (Bucher and Saenger, 1991; Digby et al., 1999; Edgar et al., 2000; Murray et al., 2006). Along the coast of Victoria, Australia, which forms the study region for this project, IOCE are common and show considerable diversity in form and entrance condition (Kench, 1999). Site specific and broad geomorphic descriptions of a small subset of IOCE in Victoria have been used to develop classification models (Arundel et al., 2008; Barton and Sherwood, 2004; Barton et al., 2008; Mondon et al., 2003). However, these classifications again focus on the ecology and local physio-chemical variations in these systems. What is needed is a classification of IOCE based on geomorphic processes that differentiates between sub-types of these estuaries from easily measurable variables of marine and fluvial parameters.

Estuary entrance condition is linked to a suite of physical factors including: (a) tidal velocity and tidal prism (Escoffier, 1940; Hinwood et al., 2012; Jarrett, 1976; O’Brien, 1931); (b) the wave climate and annual littoral drift (Bruun, 1986; Ranasinghe et al., 1999); (c) river discharge (Cooper, 1994); (d) entrance orientation (Arundel et al., 2008; Barton and Sherwood, 2004), and (e) basin accommodation space (Roy, 1984; Roy et al., 2001). For example, estuaries with larger tidal prisms, in combination with a wider and deeper entrance channel, are likely to remain open for long periods due to the more efficient scour of channel and berm sediments (Bruun, 1986; Hume and Herdendorf, 1993). Despite these broad boundary conditions being identified, there has yet to be defined a specific combination of variables that controls the diversity in entrance closure regimes of IOCE. Identifying how these catchment-scale variables act as geomorphic controls on entrance condition is important for informing future management decisions regarding IOCE.

This study aims to use catchment-scale geomorphic variables to statistically delineate IOCE and identify potential subsets within IOCE. The microtidal coast of Victoria, Australia, containing over 100 estuaries, is the principle site of investigation. Remotely sensed and field data, in combination with records of estuary entrance condition, are used to identify the geomorphic variables that are key controls on the diversity in entrance functioning. A classification model of IOCE types is developed using 35 IOCE in Victoria, Australia, with detailed records of entrance condition, and then applied to all estuaries in Victoria. Critically, as IOCE are currently managed as a single estuarine type, it is important to determine whether they are simply an end member state or a diverse form of estuary with a range of sub-types. The geomorphic based classification model presented in this study provides an important tool for predicting not only the typical frequency and duration of entrance closures, but also which estuaries are at the most at risk of declining water quality, circulation and ecological condition based on their typical closure regime. It is also an important tool that is applicable on other wave-dominated coastlines internationally where IOCE are present.

II Methods

This study maps the morphology of all of the 111 estuaries along the coastline of Victoria. Several catchment-scale geomorphic variables were measured at each site using remotely sensed data, including: entrance channel width at the mouth, estuary perimeter, estuary surface water area, catchment area, estuary length, beach width and tidal prism (Table 1). A subset of 35 IOCE with detailed entrance condition data were selected for statistical analysis. For these 35 estuaries, the beach berm elevation was surveyed in the field. The characteristics of these 35 estuaries were then used to develop a classification of IOCE based on a statistical analysis of the catchment-scale controls of the diversity in morphology and entrance function. This model was then applied to all 111 estuaries in Victoria.

Table 1.

Definition of parameters and methodology of measurement for data used to classify estuaries in Victoria. For all estuaries in Victoria, the entrance channel width, estuary perimeter, estuary length, beach width, estuary surface water area, catchment area and tidal prism were measured using remote sensing methods. Mean annual rainfall was also included. For a subset of 35 IOCE with entrance condition data, the entrance channel and berm elevation were also measured by field surveying.

Parameter	Code	Unit	Definition	Data source
Channel width	CW	m	Width of channel across lowest section of mouth	Field surveying for 35 study sites, LiDAR
Estuary perimeter	EP	km	Distance around estuary water area which is from the mouth to upper tidal limit	LiDAR
Estuary surface water area	SWA	km²	Surface water of estuary which is estuarine in nature. From mouth to upper tidal limit (for ICE this was cross-checked with values in literature as the water area taken at one point in time is dependent on entrance behavior)	LiDAR (polygons), aerial imagery, cross-checked with literature: Digby et al., 1999; Blake et al., 2000; Mondon et al., 2003; Barton and Sherwood, 2004
Berm height	BH	m ± MSL	Height of estuary mouth berm	Field surveying
Catchment area	CA	km²	Area of drainage area contributing to estuary	LiDAR, catchment delineations, cross-checked with literature (sources as per in surface water area)
Estuary length	EL	km	Length of mouth to upper tidal limit. Taken tracing the estuary thalweg to upper limit of tidal influence	LiDAR, cross-checked with literature (sources as per in surface water area)
Beach width	BW	m	Width of beach from dune to swash	LiDAR
Tidal prism	TP	x10⁶ m³	Intertidal volume of water volume of water exchanged through an estuary between MLWS and MHWS after Barber (2003)	LiDAR (polygons), aerial imagery, cross-checked with literature (sources as per in surface water area), tidal range from (BOM, 2016)
Mean annual rainfall	MAR	mm	Average annual rainfall	BOM, 2016 historic rainfall records
% time open per year	PTO	%	The percentage of time the estuary mouth is open per year	Estuary Watch data, local databases*
Mean closure duration	MCD	Days	Duration of period of individual entrance closure	Estuary Watch data, local databases*
Mean opening duration	MOD	Days	Duration of period of individual entrance opening	Estuary Watch data, local databases*
Mean closure frequency	MCF	T/yr	How many estuary entrance openings occur on average per year	Estuary Watch data, local databases*

*Lake Tyers: http://www.laketyersbeach.net.au/openings.html

Using aerial photography from 1946–2010 from the Victorian State Library, supplemented with digital globes (GoogleEarth and DigiGlobe), the entrance condition of all estuaries in Victoria was firstly classified as being either (a) permanently open or (b) intermittently closed. An estuary was classified as being open when there was a free and ongoing exchange of sediment and water between the estuarine lagoon and the ocean, and closed when a continuous subaerial berm was present at the entrance. Any estuary that was observed to have been both closed and opened at any time in the past 70 years was classified as an IOCE, as per the temporal extent of aerial imagery. Google Earth imagery was available from 1984–2016 for all estuaries in Victoria, and had an average time interval between images of six months. The total amount of historic Google Earth images available for all estuaries was 15 on average (± 8 images).

For each estuary in Victoria, measurements of the entrance channel width at the mouth, estuary perimeter, estuary surface water area, catchment area, estuary length, beach width and tidal prism were undertaken using airborne light detection and ranging (LiDAR) data, which was collected in 2007–2009 by the Victorian State Government (Table 1; Figure 1). The LiDAR survey was conducted using a LADS Mk II system coupled with a GEC-Marconi FIN3110 inertial motion sensing system and a dual frequency kinematic global positioning system (kGPS). This dataset was processed to produce a seamless terrestrial–marine 2.50 m resolution mosaic from elevations of +10 m to depths of −25 m (Quadros and Rigby, 2010) and a 1 m resolution dataset for the upper catchment with > 4 points per m². Estuary tidal prisms were calculated as per Barber (2003) with the spring tidal range taken from the Bureau of Meteorology (BOM). Obtaining field measurements of tidal current velocity and cross-sectional area for all sites state-wide was outside the scope of this study. The river flow per tidal cycle was added to the tidal prism for permanently open estuaries with large rivers (as per Ranasinghe et al., 2013), however for IOCE, due to the absence of flow gauging stations on 75% of sites, the tidal prism provides a minimum estimate as river flow was not added. Mean annual rainfall was taken from BOM with all stations having > 40 years of data and being within 10 km of each estuary. Mean annual rainfall was then analyzed relative to catchment area in order to provide an indication of the likely inflow at each estuary due to many sites not having flow gauging data.

The entrances of a subset of 35 estuaries were mapped in detail and surveyed in the field between 2012 and 2015 using a Sokkia S5X total station and Trimble R6 Real Time Kinematic (RTK) GPS. These were estuaries where entrance condition monitoring was undertaken by Estuary Watch Victoria (spanning six years on average, range 2–11 years). The entrance monitoring data included a record of the mouth status (open or closed) with supporting photographs for each time the site was visited and was supplemented with field observations during 2012–2015 and any available local databases (Table 1). The closure and opening duration of each site was obtained from taking the duration (in days) of each individual opening or closure event. Durations were totaled over the whole monitoring record for each site and then calculated as mean values. Opening and closure frequency were also calculated as an average over the extent of the monitoring period. The closure duration was recognized as the full period in which the estuary was separated from the ocean by a subaerial berm with the frequency being each individual instance of closure/opening. It is recognized that the infrequency (i.e. monthly vs weekly monitoring) of entrance monitoring at some sites may provide a minimum estimation of closure and opening frequency.

Statistical analysis of all geomorphic variables and entrance condition data was undertaken in IBM SPSS Statistics 23 and MATLAB R2015b. For the subset of 35 IOCE with records of entrance condition, all variables were tested for normality using the Kolmogorov-Smirnov coefficient and were below the threshold for normal distribution. Therefore, non-parametric methods were used to assess the relationships between variables and to classify IOCE into sub-types. A Pearson correlation analysis was performed to identify the correlations between catchment-scale variables and their influence on estuary entrance condition using the entrance condition data obtained from Estuary Watch Victoria (Table 1). Non-metric multidimensional scaling (nMDS) and hierarchal cluster analysis were undertaken using methods of Everitt and Dunn (1991); Clarke and Warwick (1994) and Saintilan (2004), to identify possible groupings between IOCE. Hierarchal cluster analysis was performed following nMDS to confirm the groupings identified using nMDS and to further identify any outliers. Following the groupings of IOCE which became apparent during nMDS and hierarchal cluster analysis, the distribution of all catchment-scale variables and entrance condition data (Table 1) were presented as boxplots for each group with descriptive statistics calculated within each group.

Multivariable linear regression using a stepwise backwards selection criteria was used to produce an equation to predict the likely percentage of time (PTO) an IOCE will exist in an open state annually. Data was log transformed to ensure a normal distribution and all measured geomorphic parameters (Table 1) for the subset of 35 IOCE were input into the multivariable linear regression to identify the strongest predictors of the independent variable (percentage of time open annually).

III Regional setting

The coast of Victoria, Australia, is 1700 km long and is divided into estuaries to the west and east of Wilson’s Promontory (Figure 2). On the western coast, mean annual significant wave height is 2.49 m with a mean annual maximum wave height of 4.13 m, mean annual wave period of 7.65 s and direction of 212.80° at Port Campbell (Hughes and Heap, 2010; Water Technology, 2004). The spring tidal range at Lorne, in the central western coast, is 1.60 m, while in the far western coast, the spring tidal range is 0.90 m at Portland. In west Victoria, south and southwesterly winds are dominant for most of the year; however, during spring and summer, winds from the south and southeast increase in frequency. This results in a predominantly easterly longshore drift for about nine months annually with a westerly occurring drift over summer (Barton and Sherwood, 2004). In eastern Victoria, (Figure 2), the mean annual significant wave height is 1.03 m with a mean maximum wave height of 1.73 m and period of 8–10 s (Short, 1996). The dominant direction of wave approach is from the south and southeast (Water Technology, 2004) with some influence from the northeast associated with east coast lows (25% of annual swell frequency) (Bird, 1993; Short, 1996). This results in a dominant longshore drift towards the northeast however the direction of drifting is reversed when north-easterly winds dominate (Bird, 1961; Short, 2010). The spring tidal range at Lakes Entrance, in the central eastern coast is 0.95 m. Wave heights peak in June to August along both coastlines.

The coastal climate of Victoria is temperate with median annual precipitation ranging between 900–1500 mm and peak rainfall occurring in the mountainous regions, particularly in the Otway Ranges (> 1400 mm/yr). Rainfall is highest in late winter to early spring and lowest in summer, and the river flow of catchments draining the coast also exhibit high inter-annual variability (Peel et al., 2001). The current coastline is tectonically stable (Murray-Wallace and Belperio, 1991) and the Holocene sea level is considered to be primarily driven by eustatic processes, with the Holocene highstand being ∼2.30 m above present sea level (Bryant, 1992).

IV Results

4.1 Geomorphic diversity of estuaries in Victoria

There are 111 estuaries in Victoria of which 103 (93%) are IOCE, having been both closed and open to the ocean at least once between 1946–2016. IOCE in Victoria typically have a shallow (< 2 m deep) lagoon impounded by a subaerial berm (Figure 3). The catchments behind IOCE are on average, < 1000 km², have entrance channels < 500 m wide and small backing estuarine dimensions (e.g. lengths < 10 km, perimeters < 10 km, surface water areas < 10 km²; Table 2). The small lagoon size leads to a mean tidal prism for these estuaries of 5.01 × 10⁶ m³ (Table 2).

Table 2.

Descriptive statistics for geomorphic parameters of all estuaries in Victoria (n = 111), all permanently open estuaries (non-modified entrances) (n = 4) and IOCE (n = 103). Parameters measured using remotely sensed data with mean annual rainfall from BOM (2016). Data parameters as per Table 1 and include: (CW) entrance channel width (m), (CA) catchment area (km²), (EP) perimeter (km), (EL) estuary length (km), (SWA) estuary surface water area (km²), (BW) beach width (m), mean annual rainfall (MAR), (TP) tidal prism (x 10⁶ m³).

All estuaries
Parameters	CW (m)	CA (km²)	EP (km)	EL (km)	SWA (km²)	BW (m)	MAR (mm)	TP (x 10⁶ m³)
Mean	146.50	711.10	6.61	4.50	4.63	93.30	951	7.19
Std dev	218.10	1967	9.01	5.73	9.53	56.34	227	9.01
Median	66.50	125	2.58	2	0.94	89.50	951	2.34
Minimum	34	14	0.70	0.10	0.01	25	600	0.05
Maximum	2100	8590	42	19	30	340	1431	29.80
Permanently open estuaries (non-modified entrances)
Mean	1155	3321	28.80	16.86	21.44	198.75	891	26.38
Std dev	682.62	3523	14.27	2.77	2.04	100.53	115	2.62
Median	1020	1597	33	17.83	20.30	175	891	26.16
Minimum	480	1500	8.20	12.80	15.20	105	721	24.30
Maximum	2100	8590	42	19	30	340	976	29.80
Intermittently open/closed estuaries (all IOCE in Victoria)
Mean	120	499	5.47	1.74	3.30	73.30	964	5.01
Std dev	177	432	6.1	4.43	4.46	66.80	220	6.54
Median	62.40	409	2.48	1.74	0.95	110	951	1.92
Minimum	34	14	0.07	0.10	0.01	25	600	0.05
Maximum	600	1110	22	17	19	187	1431	22.50

On the west coast, there are 46 IOCE, the largest being Curdies Inlet with a channel width of 314 m and a catchment area of 1110 km². The estuary has a large central basin (12 km²) which is up to 2 m deep. In contrast, the smallest IOCE is Moggs Creek (channel width of 34 m, surface water area < 1 km²). This estuary is found at the base of a steep catchment of 14 km², which forms in the Otways Ranges and exits onto the beach via a channel < 0.25 m deep. The Fitzroy River could be considered to have a typical morphology for a west coast IOCE (Figure 3(e)). It has an entrance channel width of 160 m, a catchment area of 260 km² and an estuarine lagoon of 1.20 km² which is, on average, 1.25 m deep.

On the east coast, 57 IOCE are present with a similar range of dimensions to those found on the west coast (Table 2). The largest IOCE is Sydenham Inlet with a channel width of 600 m, perimeter of 22 km and a surface water area of 19 km². The estuary has a large central basin (22 km²) which is up to 3 m deep. The smallest IOCE in east Victoria is Davis Creek, which has an entrance channel width of 35 m, a catchment area of 15 km² and a surface water area < 1 km². There are numerous small IOCE (40% of all east Victorian IOCE) around the granitic Wilsons Promontory, occupying steep, small catchments with areas < 30 km², channel widths < 50 m and surface water areas < 1 km² (e.g. Figure 3(f)).

Of the eight permanently open systems present on the open coast of Victoria, four estuaries had unmodified entrances (e.g. Barwon River, Andersons Inlet, Shallow Inlet, Mallacoota Inlet) with the other half being kept open through human intervention (e.g. entrance training walls at Moyne River). Permanently open estuaries tend to have large catchments (> 1000 km²) with entrance channel widths (> 450 m) and basin dimensions all above the mean values for estuaries in Victoria (Table 2). For example, the Barwon River in west Victoria (Figure 3(a)) has a catchment area of 8590 km², a length of 19 km, a depth in the central basin of > 3 m and an entrance channel 480 m wide. Similarly, Andersons Inlet in east Victoria (Figure 3(b)) has a length of 17.65 km, a channel width of 2 km, catchment area of 1600 km² and a surface water area of 13 km² (Table 2). Permanently open estuaries in Victoria without entrance modification had tidal prisms ranging between 23.40 and 29.80 × 10⁶ m³ (Table 2) and tended to have mouths which were orientated away from the dominant wave direction (Table 3).

Table 3.

Entrance configuration and entrance condition of a subsample of 35 IOCE and four permanently open estuaries without entrance modification relative to the dominant direction of wave approach on both the east and west coasts of Victoria. PTO refers to the percentage of time open annually.

		East coast	West coast
Dominant wave direction		SE	SW
Number of sample IOCE	35	13	22
Number of IOCE with entrances facing:	SE	7	2
	S	6	4
	SW	0	11
	W	0	5
Mean annual PTO (%) for of IOCE with entrances facing:	SE	56	72
	S	44	88
	SW	-	69
	W	-	55
Dominant mouth orientation of IOCE		SE	SW
Number of sample permanently open barrier estuaries	4	2	2
Number of permanently open barrier estuaries with entrances facing:	SE	1	1
	S	1	1
	SW	0	0
Dominant mouth orientation of permanently open barrier estuaries		S	S–SE

4.2 Controls on entrance condition and estuary morphology

A continuum of catchment, channel and estuary basin sizes is evident across the 35 field-surveyed IOCE in Victoria showing that the estuary dimensions of these IOCE are positively scaled to catchment area and channel width (Figure 4; Table 4; Table 5). Using a Pearson correlation analysis, the relationship between the catchment-scale variables measured for the 35 field-surveyed IOCE were tested for correlations to determine relationships between the dimensions of the estuary basin, channel and catchment and their influence on entrance functioning. The highest positive correlations occur between tidal prism and surface water area (p = 0.981, γ = 0.01 level), estuary length and perimeter (p = 0.923, γ = 0.01 level), estuary perimeter and channel width (p = 0.921, γ = 0.01 level), estuary perimeter and surface water area (p = 0.906, γ = 0.01 level), perimeter and tidal prism (p = 0.902, γ = 0.01 level), and catchment area and tidal prism (p = 0.901, γ = 0.01 level) (Figure 4; Table 5). Berm height also showed a positive relationship with all other geomorphic variables (p = 0.425– 0.555, γ = 0.01 level) (Table 5).

Table 4.

Descriptive statistics for geomorphic and entrance parameters of the 35 case study IOCE sites in Victoria. Data is derived from sites which were visited and surveyed as part of field work, and with entrance condition monitoring data. Entrance condition monitoring recorded by Estuary Watch Victoria online database system and field observations.

Parameter	Minimum	Maximum	Mean	Std dev
Channel width (m)	34	600	139	127
Catchment area (km²)	14	1200	308	347
Surface water area (km²)	0.01	19	5.90	9.70
Estuary perimeter (km)	0.70	22	7.90	7.60
Estuary length (km)	0.50	17	5.20	4.50
Berm height (m above MSL)	0.10	3.20	1.30	0.90
Tidal prism (x 10⁶ m³)	0.05	22.50	5.01	6.54
Mean annual rainfall (mm)	600	1431	964	220
Percentage of time open annually (%)	18	98	60.90	31.40
Mean opening frequency (T/yr)	0.50	10	2.70	2.40
Mean closure duration (days)	1.30	372	64.70	99.70
Mean opening duration (days)	14	450	185	126.60

Table 5.

Pearson correlation analysis of geomorphic parameters and their relation to IOCE morphology and entrance condition. Parameters as per Table 1 and data for 35 case study IOCE with detailed entrance condition data (n = 35 for each parameter).

		CW	CA	TP	EL	BH	SWA	PTO	MCF	MCD	TP	MAR
CW	Pearson correlation	1	0.859**	0.921**	0.885**	0.425*	0.798**	−0.670**	−0.176	0.724**	0.788**	−0.286
	Sig. (2-tailed)		0	0	0	0.011	0	0	0.311	0	0	0.096
CA	Pearson correlation	0.859**	1	0.883**	0.892**	0.471**	0.875**	−0.603**	−0.08	0.681**	0.901**	−0.375*
	Sig. (2-tailed)	0		0	0	0.004	0	0	0.649	0	0	0.026
P	Pearson correlation	0.921**	0.883**	1	0.923**	0.555**	0.906**	−.722**	−0.066	0.768**	0.902**	−0.364*
	Sig. (2-tailed)	0	0		0	0.001	0	0	0.708	0	0	0.032
EL	Pearson correlation	0.885**	0.892**	0.923**	1	0.477**	0.845**	−0.653**	−0.039	0.704**	0.847**	−0.363*
	Sig. (2-tailed)	0	0	0		0.004	0	0	0.823	0	0	0.032
BH	Pearson correlation	0.425*	0.471**	0.555**	0.477**	1	0.489**	−.649**	0.571**	0.634**	0.497**	−0.477**
	Sig. (2-tailed)	0.011	0.004	0.001	0.004		0.003	0	0	0	0.002	0.004
SWA	Pearson correlation	0.798**	0.875**	0.906**	0.845**	0.489**	1	−0.628**	−0.028	0.674**	0.981**	−0.203
	Sig. (2-tailed)	0	0	0	0	0.003		0	0.875	0	0	0.242
PTO	Pearson correlation	−0.670**	−0.603**	−0.722**	−0.653**	−0.649**	−0.628**	1	−0.216	−0.949**	−0.603**	0.432**
	Sig. (2-tailed)	0	0	0	0	0	0		0.212	0	0	0.01
MCF	Pearson correlation	−0.176	−0.08	−0.066	−0.039	0.571**	−0.028	−0.216	1	0.211	−0.033	−0.404*
	Sig. (2-tailed)	0.311	0.649	0.708	0.823	0	0.875	0.212		0.224	0.851	0.016
ACD	Pearson correlation	0.724**	0.681**	0.768**	0.704**	0.634**	0.674**	−0.949**	0.211	1	0.650**	−0.454**
	Sig. (2-tailed)	0	0	0	0	0	0	0	0.224		0	0.006
TP	Pearson correlation	0.788**	0.901**	0.902**	0.847**	0.497**	0.981**	−0.603**	−0.033	0.650**	1	−0.249
	Sig. (2-tailed)	0	0	0	0	0.002	0	0	0.851	0		0.149
MAR	Pearson correlation	−0.286	−0.375*	−0.364*	−0.363*	−0.477**	−0.203	0.432**	−0.404*	−0.454**	−0.249	1
	Sig. (2-tailed)	0.096	0.026	0.032	0.032	0.004	0.242	0.01	0.016	0.006	0.149

*Correlation is significant at the 0.05 level (2-tailed).

**Correlation is significant at the 0.01 level (2-tailed).

4.3 Dynamics of estuary mouths

As channel width at the estuary entrance has a consistently strong positive relationship with all other geomorphic variables, this was used to illustrate the influence of all other scaled geomorphic parameters in relation to entrance functioning (Figure 4(a)–(i)). The 35 field-surveyed IOCE closed between 1 and 10 times annually (mean 2.70 ± 2.40) (Table 4) with the lowest closure frequency occurring on smaller estuaries (channel widths < 75 m; catchments < 75 km²) (e.g. Kennett River, Skenes Creek and Wye River). The mean closure frequency showed a negative relationship with all geomorphic parameters aside from berm height (p = 0.571, γ = 0.01 level) (Table 5). The total percentage of time open per year was highly variable ranging between 18% and 98% (mean 61% ± 31%) with the highest percentage of time open occurring at Wild Dog Creek and Cumberland River, both with small channel, basin and catchment dimensions (Figure 4 (g)). The lowest proportion of time open was at Curdies Inlet, the largest IOCE in western Victoria (18%). Surprisingly, the percentage of time open annually showed a strong negative relationship with all geomorphic variables (Table 5). The mean closure duration was also variable and ranged between one and 372 days (mean of 64 days). Closure duration showed a strong positive relationship with all geomorphic variables (p = 0.768–0.634, γ = 0.01) (Table 5). The longest closure duration occurred at larger IOCE such as Curdies Inlet, while IOCE with the smallest entrance channels and dimensions have the shortest closure durations, that persisted for days on average. Additionally, the mean duration of each entrance opening was 184.90 ± 126.60 days with the maximum occurring at both the larger and smaller end member IOCE (Table 4).

All surveyed estuary mouths had a broad U-shaped cross section. At the mouth, the base of the widest channels (> 200 m) was < 0.50 m below MSL (Figure 5a). In long profile, the berm crest elevation was +0.5–1 m above MSL with a landward decrease in elevation to below MSL upon entering the lagoon (Figure 5b). In channels between 75 and 200 m wide, the channel bed depth at the mouth was generally above MSL. Berms were steep rising to > +1 m above MSL (Figure 5b). The base of the narrowest channels (< 75 m) were generally the deepest, incising to < −1 m below MSL where the thalweg was present (Figure 5c). In long profile, a distinct berm was not evident in channels of < 75 m size (Figure 5c). The long profile morphology may, however, be more a function of the entrance conditions at the time of surveying because all IOCE with channel widths < 75 m were open and surveyed in their dominant entrance state. At the time of surveying, all other IOCE were closed, so closed profiles are used for comparison.

4.4 Classification of IOCE in Victoria

Using a non-metric multidimensional scaling analysis (nMDS), three distinct groupings of IOCE are apparent (Figure 6). The 2D stress factor for nMDS was 0.01 and a value of < 0.1 indicates a significant relationship between variables (Clarke and Warwick, 1994). The first group includes the larger estuaries such as Curdies Inlet and Lake Tyers, which have entrance channel widths > 200 m, catchment area > 500 km² and tidal prism > 10 × 10⁶ m³ (Figure 6, Group 1). The entrances of this group are open less frequently, but for longer periods of time. In the second group, which includes estuaries of mid-size range (75–200 m channel width, 75–1000 km² catchment area and tidal prism 1.50–10 × 10⁶ m³), the entrance condition is highly variable. Entrances are open for 19–90% of the time annually with 1–10 openings per year. Examples in Victoria include Merriman Creek and Aire River (Figure 6, Group 2). Within this group, the nMDS plot shows that the Gellibrand River could be considered an outlier. This is due to the larger basin and catchment dimensions (i.e. catchment > 1000 km², perimeter > 10 km and length > 10 km) but lower surface water area as attributed to the potentially higher level of sedimentary infill present. The final group is most tightly clustered and includes the smallest IOCE with entrance channel widths < 75 m, catchments < 75 km² and tidal prisms < 1.5 × 10⁶ m³ (Figure 6, Group 3).

Hierarchical cluster analysis produces similar results to nMDS (Figure 7). The larger estuaries of Sydenham, Wingan, Tamboon and Curdies Inlets and Lakes Tyers and Bunga are clustered together (Figure 7, Group 1). Group 1 estuaries can be further subdivided between Lakes Bunga and Tyers, Curdies Inlet, Wingan Inlet and Sydenham and Tamboon Inlets as based on the differences in basin and catchment size (i.e. Sydenham and Tamboon Inlets are most similar in surface water area and have the largest catchments > 1100 km²; Figure 7). As per nMDS analysis, the Gellibrand River is considered an outlier. Relative to all other IOCE, however, dimensions of all Group 1 estuaries were an order of magnitude larger than the smallest IOCE (Group 3) and > 5 km² larger than the next highest value in Group 2. The next most distinct cluster (Group 2: groups B1, B2, B3) subdivide between Thurra River and Sydenham Inlet (Figure 7).

Group 2 represents the mid-sized IOCE which have channel widths 75–200 m and the most variable entrance condition. Within Group 2 there is a secondary divergence, which is likely attributed to the varying degrees of basin accommodation space associated with these IOCE, ranging between the two end member types. Group B3 (Group 3) represents the smaller IOCE with channel widths < 75 m (Figure 7). Compared to nMDS results, the dendrogram shows more overlap between groups. This is explained by the split between Mueller River (B2) and Shipwreck Creek (B3) being predominantly based on estuary and catchment size, as opposed to recognizing the different entrance conditions and topographic setting that may otherwise differentiate between these estuaries. For example, Cumberland, Wye, Kennett and St George rivers are open for > 90% of the time annually and have limited to no central basin accommodation space, but have catchments ranging between 20 and 36 km². Barnham River and Aireys Inlet, also included in Group B3, are both extensively infilled and open for < 60% of time annually; however, because they have catchments 79–81 km², they too classify as Group 3 IOCE. As a result, Group B3 is included in Group 3 in the nMDS plot, the smaller end member of all IOCE but with some overlap as hierarchal analysis does not account for infill and topographic setting, and appears to prioritize spatial scale over the dominant entrance condition.

The distribution of variables within these three groupings of IOCE further confirms the nMDS and hierarchal cluster results with distinct differences between the distribution and median values for each group (Figure 8a–l). It is evident that Group 1 consistently show the largest channel, catchment and basin dimensions, Group 2 occupy the mid-range of scale but have the most variable range of values, and Group 3 are the smallest, with the most tightly distributed values on average (Figure 8a–k). When considering the impact of these groupings on entrance condition, Group 1 close to the ocean less often; however, entrance closure and opening both persist for months at a time (Figure 8i–l). Group 2 are highly variable in terms of both opening and closure frequency and duration, and Group 3 are open for the majority of the time, with closures only persisting for days (Figure 8i–l).

4.5 Predicting entrance condition in IOCE

Multivariable linear regression was used to produce an equation to predict the likely percentage of time an IOCE will exist in an open state annually prior to any human modification to the catchment or entrance morphology (equation (1)). Channel width (CW), catchment area (CA) and surface water area (SWA) were found to be the strongest drivers of the annual percentage of time open (PTO) in IOCE. Multivariate analysis had a standard error of +0.19, degrees of freedom (df) value of 30 and a R Square value of 0.61 (i.e. 61% of variation in the response variable: PTO is predicted using these parameters). PTO can therefore be accounted for using the three predictor values (CW, CA and SWA). As a result, equation (1) is used to model PTO in IOCE in Victoria where CW is channel width, CA is catchment area and SWA is surface water area

\begin{array}{l} LogPTO = 2.424 + (- 0.686) (logCW) \\ + (0.323) (logCA) \\ + (- 0.250) (logSWA) \end{array}

Channel width, as a predictive parameter represents the capacity of the entrance channel to facilitate outflow and exchange with the ocean. Surface water area relative to catchment area is effectively a proxy for the amount of basin accommodation space along with representing the size of the backing lagoon as a whole. Catchment area provides an approximation of how much fluvial input into the estuary basin occurs in terms of runoff, discharge and groundwater flow.

V Discussion

5.1 Classification model of IOCE in Victoria

93% of estuaries in Victoria close intermittently; therefore, this is a fundamental property of Victorian estuaries. Of the small number of permanently open systems, most remain open due to human intervention through features designed to counteract infilling of the entrance channel. The larger tidal prism and entrance channels at those sites which remain permanently open without entrance modification reflects the well-known AP empirical stability relationship where a larger entrance cross-sectional area (A) and tidal prism (P) corresponds with a more stable channel (Hume and Herdendorf, 1993; Jarrett, 1976; O’Brien, 1931). As wave energy along the Victorian coast is persistently high due to the influence of uninhibited swell propagation from the Southern Ocean, littoral drift would also be high by global standards (Bruun, 1986; Davis, 1989). For example, median annual wave power exceeds 30 kW/m in west Victoria with a mean annual H_s > 2 m (Hughes and Heap, 2010; Water Technology, 2004). While these sites may experience similar wave conditions to IOCE, the larger tidal prisms mean that a higher proportion of sediment can be removed from the entrance channel due to a larger volume of tidal exchange and higher ebb-tidal velocities (Bruun, 1978; Bruun and Gerritsen, 1960). The larger catchments also add more river flow, thus strengthening the ebb-tidal prism (Table 2). In Victoria, these sites also tend to have entrances which do not directly face the dominant direction of wave approach, meaning local topographic features may shelter these sites from the direct influence of waves (Table 3).

Within IOCE in Victoria, the continuum of entrance channel width, basin dimensions (perimeter, basin surface water area, estuary length), tidal prism and catchment size can be used as a basis for classifying IOCE (Table 6; Figure 9). Statistical analysis indicates that three clear sub-types of IOCE can be identified in Victoria: Type A – large IOCE (corresponding to Group 1); Type B – medium sized IOCE (Group 2); and Type C – small, tidal creeks (Group 3) (Figure 6; Figure 7; Figure 9). As the variability in these geomorphic parameters in turn controls entrance functioning, each sub-type of IOCE exhibits a different characteristic opening and closure regime. For example, there is at least one order of magnitude difference between mean closure duration in the three types of IOCE.

Table 6.

Parameters evident in Victoria for conceptual classification of IOCE groups. The top line for each group indicates the thresholds for the classification for estuaries in Victoria. Parameters are taken from 35 case study IOCE sites and are as defined in Table 1 including: entrance channel width (CW), catchment area (CA), estuary surface water area (SWA), estuary perimeter (km) (EP), estuary length (EL), tidal prism (TP), mean annual rainfall (MAR), mean closure duration (MCD) and percentage of time open annually (PTO).

IOCE type	CW (m)	CA (km²)	SWA (km²)	EP (km)	EL (km)	TP (x10⁶m³)	BH (m + MSL)	MAR (mm)	MCD (days)	PTO (%)
Group 1 (Type A)	> 200	> 500	> 5	> 12	> 10	> 10	1–2	800–1000	> 150	< 35
Mean	265	762.5	17.8	20	12.6	15.5	1.72	892	305	32
Median	321	925	16.5	21	15	15.1	1.7	891	304	15
Range	203–600	500–1200	15–22	15.3–25	8–17	7–19.8	2–1.6	826–967	290–372	18–80
Std dev	60	240	11.80	3.8	4.1	3.9	0.2	64.5	10	28.2
Group 2 (Type B)	75–200	75– 000	1–10	2–12	2–10	1.5–10	1.5–4	600–1100	15–200	20–90
Mean	120	327	4.2	7	5.1	5	1.9	822	145	49
Median	124.1	305	1.9	7.5	6.1	2.6	1.8	895	172	50
Range	101–141	68–1000	1.1–15	3.3–11.8	2–10	1.6–19.5	3.2–1.3	600–1070	11–234	19–90
Std dev	18	278	5.4	2.7	2.6	5.4	0.8	170.2	64	22
Group 3 (Type C)	< 75	< 75	< 1	< 2	< 3	< 1.5	0–0.3	> 1000	1–15	≥ 90
Mean	42.9	32	0.4	2	1.4	0.7	0.4	1134	10	93
Median	44	29.2	0.2	2.2	1.3	0.2	0.1	1100	25	95
Range	34–66	14–75	0.1–1	0.7–2.2	0.6–3.1	0.1–2.1	0–1.2	944–1431	1–17	90–98
Std dev	12	18	0.4	1.2	0.7	0.7	0.8	194.7	6	4

5.2 Type A – large IOCE

Type A IOCE exhibit the largest estuary dimensions in terms of channel width (> 200 m), estuary water surface area (> 5 km²), length (> 10 km), perimeter (> 12 km), catchment size (> 500 km²) and tidal prism (> 10 × 10⁶ m³) (Table 6; Figure 8; Figure 10). This increase in catchment size, basin water surface area and tidal prism (Figure 8) is consistent with past literature (Haines et al., 2006; Mondon et al., 2003; West et al., 1985). Type A estuaries form on low-gradient coastal plains and have a basin morphology which tends to be circular, with a large central basin and ample accommodation space. They have a distinct beach berm at an elevation (when closed) of +0.50–1.50 m MSL (Figure 5a).

Type A IOCE experience few openings per year (1–2 per year) being open on average for 32% of the time annually; however, they are distinguished by long periods of both opening and closing (Table 6). When entrance openings occur, they persist for more than a month, while closures typically last for months to years at a time. For example, on average, Curdies Inlet was closed for just 18% of the time annually, but had a mean opening duration of 200 days and mean closure duration of 210 days. The monthly scale duration of entrance opening is due to the larger tidal prism and increased hydraulic efficiency through a larger entrance channel (Bruun, 1978; Gao and Collins, 1994; O’Brien, 1931). Evidence of this is the deposition of prominent flood tidal deltas indicating periods of sustained tidal exchange during past openings (Figure 10a). As the lagoon drains, and in the absence of sufficient fluvial flow to add to the ebb-tidal prism, the high-wave energy of the Victorian coastline is likely to close these larger IOCE faster than if they were located on coasts with a less energetic wave climate and lower drift rate. The frequency and duration of entrance opening and closure is therefore not only dictated by contemporary processes such as the relative balance of wave and fluvial energy, but also largely by the underlying basin topography and relative accommodation space. This is a fundamental scaling feature of IOCE in Victoria as, while the marine conditions are similar across the state, those estuaries with larger catchments and lagoons can maintain an open entrance for longer due to the larger tidal prism. Once isolated from the ocean and tidal influence, however, the larger basin volume takes longer to fill, until overtopping from the catchment side reinitiates opening.

In Australia, Type A IOCE are a counterpart of the large, immature ‘saline coastal lagoons’ found in New South Wales (NSW) (Roy, 1984; Roy et al., 2001 – Type IV 8). These sites have similar morphologies; for example, Smiths Lake (extensive tidal delta deposits, channel width > 200 m, surface water area 10 km²) and Coila Lake (surface water area 7 km², closed 95% of time) (Roy et al., 2001). A fundamental difference between these sites and Victorian Type A IOCE is that in NSW these IOCE showed large basin and channel dimensions, but often with small catchments; for example, Smiths Lake (33 km²) and Coila Lake (52 km²). As a larger catchment area has a positive influence on the percentage of time open annually (equation (1)), Type A IOCE with small catchments would be expected to be closed for a higher proportion of the time compared to those with larger catchments, as their basin size is similar but there is less inflow to the estuary. This underpins the role of estuary basin accommodation space as a key control on opening frequency and duration. Internationally, Type A IOCE may also be considered a counterpart of the ‘Waituna’ coastal lakes found in New Zealand, such as Waituna Lagoon and Lakes Ellsemere and Forsyth, but forming on sandy beaches. Waituna have large surface water areas (> 5 km²), are typically closed for months to years at a time and are associated with low river flows relative to catchment size (Kirk and Lauder, 2000; NIWA, 2016; Schallenberg et al., 2010). The large ‘coastal lagoons’ of northern California are another international form of Type A IOCE, including Stone Lagoon and Lake Earl (Kraus et al., 2008). These sites have catchments > 500 km², entrance channels > 200 m, tidal prisms > 5 × 10⁶ m³ and close every 1–2 years for months to years at a time (Kraus et al., 2002, 2008). Consistent with Victoria, both the national and international counterparts of Type A IOCE exist in a predominantly closed entrance state, with closures persisting for months to years.

5.3 Type B – medium sized IOCE

Type B IOCE occupy the mid-spectrum in terms of size, and are smaller than Type A, with channel widths 75–200 m, tidal prisms 1.50–10 × 10⁶ m³, catchment areas 75–1000 km², estuary surface water areas 1–5 km², lengths 2–10 km and perimeters 2–12 km (Table 6; Figure 11). The catchment and estuary dimensions of Type B IOCE showed more variance than other types of IOCE, likely due to the wider range of coastal and topographic settings that they occupy and their varying degrees of sedimentary infill. For example, the Gellibrand River has a catchment area > 1000 km², a channel width of 120 m, a channelized (< 3 m depth) central basin with extensive infill, and a widespread fringing swamp (Barton and Sherwood, 2004). The Anglesea River, in contrast, has a catchment area of 125 km² and an entrance channel of 90 m with a shallow (< 1.50 m) central basin. Throughout the central basin of Type B IOCE, width increases laterally, but then becomes narrow again at the head of the estuary. Type B IOCE exhibit a steep beach berm, but the range of berm heights is variable due to the variety of beach types forming at the entrance. For example, both Sherbrook River and Spring Creek in west Victoria occupy rocky embayments and have berm heights of +1.50–3.15 m above MSL. In contrast, the Thurra, Aire and Hopkins estuaries enter the open coast and have berm heights of +2.20–3.05 m above MSL. Berms increase in elevation rapidly over the first 20 m landward of the swash zone where after 20 m the channel bed elevation only decreases slightly (±0.30 m) to indicate further sediment storage behind the berm (Figure 5b).

Type B IOCE entrances open at a weekly to monthly frequency (1–10 times/year) (Table 6). Entrances stay open for less time than Type A IOCE, typically for weeks to months. For example, the Aire River opens on average four times per year for a mean duration of 21 days, while the Anglesea River opens five times per year with a mean opening duration of 14 days. Entrance closures are more frequent than in Type A IOCE, but of a shorter duration as the lower volume of flood accommodation space requires less freshwater input before the berm is overtopped from the catchment side. Entrance openings also last for weeks to months as the smaller tidal prism means that marine depositional processes can dominate more rapidly in the absence of a rapid influx of river flow.

In Australia, Type B IOCE are a counterpart of a semi-mature “saline coastal lake” where extensive infill of the central basin is evident (Roy, 1984). Internationally, they are similar to the “perched, normally closed estuaries” defined by Cooper (2001), which have steep berms, an estuary bed maintained at or above MSL at the mouth, channels ∼100 m wide and openings that typically persist for weeks at a time. The Mhlanga, Mdloti, Groot Brak and East Kleinemonde rivers in South Africa are examples of estuaries which can be classified as Type B IOCE. These sites have catchments ranging between 43 and 550 km², basin surface water areas of 0.50–3 km² and entrance channels 100–200 m wide at the mouth (Cooper, 2001; Nunes and Adams, 2014; Riddin and Adams, 2008; Whitfield and Bate, 2007). These sites open between 1 - 6 times per year for a duration of weeks - days on average (Anandraj et al., 2007; Mundree et al., 2003; Wooldridge, 1994). In the USA, Type B IOCE are similar in morphology and entrance functioning to the larger “bar-built” estuaries of California (Kraus et al., 2008). Examples include the Russian River and Redwood Creek, which also have high steep berms (> +2 m MSL), similar catchment and estuary dimensions, and close several times annually for weeks to months at a time (Behrens et al., 2013).

5.4 Type C – small tidal creeks

Type C IOCE exhibit the smallest estuary dimensions with channel widths < 75 m, tidal prisms < 1 × 10⁶ m³, surface water areas < 1 km², lengths < 2 km, perimeters < 3 km and catchments < 75 km² (Table 6; Figure 12). The typical entrance morphology of Type C IOCE shows a low berm (all < +0.50 m above MSL) and smaller cross-sectional areas at the mouth (Figure 5c). Type C IOCE exhibit a narrow and meandering channel up to 200 m upstream of the mouth which, upon flowing onto the beach, becomes shallow (< 0.30 m) and fan-shaped. Tidal terraces are evident along the lower reaches of the channel, which indicate the tidal flow regularly enlarging the channel under different tidal elevations (Figures 5c). There is a generally constant depth landward of the entrance, except for deeper sections where the channel is confined by areas of bedrock or boulders (Figure 5c). Type C IOCE occupy steep, mountainous catchments with high rainfall and where the topographic setting limits the size of the estuary itself with a smaller inherited river valley.

At all study sites, Type C IOCE were open for > 90% of the year (Figure 8j). Entrance closures were also infrequent (1.18 per/ year on average) and of a short duration, persisting only over a daily or tidal scale (Figure 8g–i; Table 6). Although Type C IOCE have small basin dimensions and tidal prisms, which has previously indicated an estuary that is more prone to closure (Bruun, 1978; Bruun and Gerritsen, 1960; Jarrett, 1976), their steep, mountainous catchments with high runoff rates means that they are maintained open by more constant fluvial discharge. Type C IOCE in Victoria are located in areas with high rainfall relative to catchment size, meaning that the more constant runoff is likely to dominate over marine processes of deposition under normal conditions (Figure 8i). For example, Type C IOCE on Wilson’s Promontory (e.g. Darby River) and the Otways Ranges (e.g. the Wye and Kennet rivers) receive the highest annual rainfall state-wide of > 1400 mm (Table 6). In comparison, mean annual rainfall at Type A IOCE is between 800 and 1000 mm and at Type B IOCE is between 600 and 1050 mm (Table 6). Type C IOCE are also located seaward of coastal ranges, with a more impermeable substrate, meaning rainfall is channeled directly downslope with little opportunity to disperse through the water table. For example, the underlying granitic rock on Wilsons Promontory assists runoff from the mountainous topography (Hill, 1994). As rainfall and river flow is a main control on entrance opening, any change in precipitation or catchment characteristics will be rapidly reflected in the entrance. Many Type C IOCE in Victoria also tend to enter the ocean alongside rocky headlands, which may act to promote an area of deeper scour of the channel thalweg alongside the headland, thus extending the entrance opening duration under normal wave conditions (Figure 12a–c).

In Victoria, Type C IOCE are a counterpart of the “intermittently open small coastal creeks” described by Roy et al. (2001) occurring in NSW (Type IV 9 in their classification). A key distinction is that many small coastal creeks in NSW are infilled (i.e. mature), whereas Type C IOCE in Victoria are small on account of the topographic constraints of their confined valley setting, which limits central basin development. Internationally, Type C IOCE are a smaller counterpart of the “normally open, river-dominated” estuaries of South Africa in the sense that they are open to the ocean for the majority of the time due to the influence of river flow, lack a distinct flood tidal delta and are found in steep catchments (Cooper, 2001). Again, a key distinction between normally open, river-dominated estuaries and Type C IOCE is that, in Victoria, the confined valley setting is the primary factor limiting the estuary size as opposed to progressive sedimentary infill (Cooper, 1994; Cooper, 2001). In southern California, the “steep coastal drainage (W-C)” coastal lagoons are similar in morphology to Type C IOCE (Jacobs et al., 2011). These estuaries have small (< 75 km²), high gradient catchments and drain the face of coastal ranges, with IOCE of the Santa Lucia Range (e.g. Toro Creek) being good examples (Bell et al., 2011; Jacobs et al., 2011).

5.5 Regional distribution of estuary types

Across Victoria, permanently open estuaries are rare and the three types of IOCE identified in the classification model are relatively evenly distributed along both the east and west coasts (Figure 13a and b). Type A IOCE are less common due to their characteristic large basin area whereby these systems require an extensive past valley space to occupy. They are also found only on open coastlines. Type B IOCE occupy the most diverse topographic settings. They are present along sheltered embayments and on open coasts. Type C IOCE are concentrated in areas of high gradient catchment topography and areas with narrow coastal plains. In Victoria, there are two regions with a high proportion of Type C IOCE: the Otways coast and Wilson’s Promontory.

VI Conclusion

In Victoria, IOCE are the dominant type of estuary and constitute 93% of all estuaries along the 1700 km long open coastline. IOCE in Victoria exhibit considerable morphological diversity in terms of spatial scale, topographic setting and entrance functioning, thus comprising a suite of estuaries in their own right. This study presents a geomorphic classification of IOCE types that is both locally and internationally relevant as it enables any IOCE to be classified based on easily obtainable spatial and field data to predict entrance functioning. The delineation between IOCE types is an important advance from previous models which recognize IOCE as a homogenous estuary type despite clear differences in the frequency and duration of entrance closures. IOCE in Victoria can be classified into three distinct sub-groups: Type A (large IOCE); Type B (medium sized IOCE); and Type C (small, tidal creeks). These three types of IOCE showed an order of magnitude difference in entrance closure duration. The channel width, catchment area and lagoon size proved to be key geomorphic controls on opening and closure frequency and duration whereby a larger channel area, estuarine basin and tidal prism resulted in less frequent openings but both openings and closures of a longer duration.

Identifying the typical frequency and duration of entrance closures occurring in each type of IOCE provides a valuable asset for informing management decisions such as predicting the potential risk of each estuary type to the adverse effects of prolonged closure. By being able to predict the closure frequency and duration of IOCE, this information can be used to develop inventories of sites which would be at a high risk of water quality issues associated with longer water residence times, decreased fish passage to the ocean and flooding of surrounding land. An important consideration is that as many IOCE internationally exist in increasingly modified catchment settings where dams and irrigation practices reduce river flow into the estuary, a decrease in the current opening frequency and duration can be expected to occur in the future, with further implications for IOCE ecology and management.

Footnotes

Acknowledgements

Neville Rosengren is thanked for providing photos of estuaries and Estuary Watch Victoria is thanked for records of entrance condition. We thank the two anonymous reviewers for helpful comments on an earlier draft of this paper.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The Victorian Department of Environment, Land, Water & Planning provided financial support for this project and the first author was supported by an Australian Postgraduate Award and a Faculty of Science Postgraduate Writing up Award.

References

Allen

Posamentier

(1994) Transgressive facies and sequence architecture in mixed tide-and wave-dominated incised valleys: Example from the Gironde Estuary, France. Society of Economic Paleontologists and Mineralogists, Special Publication 51: 225.

Anandraj

Perissinotto

Nozais

(2007) A comparative study of microalgal production in a marine versus a river-dominated temporarily open/closed estuary, South Africa. Estuarine, Coastal and Shelf Science , 73(3): 768–780.

Arundel

Barton

Becker

. (2008) Reporting on the Environmental Condition of Victorian Estuaries: A discussion paper. Australia: Deakin University.

Barton

Sherwood

(2004) Estuary Opening Management in Western Victoria: An information analysis. Melbourne Victoria: Parks.

Barber

(2003) Analytical modeling of pollution flushing in well-mixed tidal embayments. Elsevier Oceanography Series 67: 467–489.

Barton

Pope

Quinn

. (2008) Identifying threats to the ecological condition of Victorian estuaries. Melbourne, Victoria: Department of Sustainability and the Environment Technical Report.

Behrens

Bombardelli

Largier

. (2013) Episodic closure of the tidal inlet at the mouth of the Russian River: A small bar-built estuary in California. Geomorphology 189: 66–80.

Bell

Dagit

Ligon

(2011) Colonization and persistence of a southern California steelhead (Oncorhynchus mykiss) population. Bulletin, Southern California Academy of Sciences, 110(1): 1–16.

Bell

KNI

Cowley

Whitfield

(2001) Seasonality in frequency of marine access to an intermittently open estuary: Implications for recruitment strategies. Estuarine, Coastal and Shelf Science, 52(3): 327–337.

10.

Bird

ECF

(1961) The coastal barriers of east Gippsland, Australia. The Geographical Journal 127(4): 460–468.

11.

Bird

ECF

(1993) The Coast of Victoria: The shaping of scenery. Melbourne: Melbourne University Press.

12.

Boyd

Dalrymple

Zaitlin

(1992) Classification of clastic coastal depositional environments. Sedimentary Geology 80(3): 139–150.

13.

Bruun

(1978) Stability of tidal inlets: Theory and engineering. Developments in Geotechnical Engineering. Amsterdam: Elsevier Scientific, 510.

14.

Bruun

(1986) Morphological and navigational aspects of tidal inlets on littoral drift shores. Journal of Coastal Research 2(2): 123–145.

15.

Bruun

Gerritsen

(1960) Stability of Coastal Inlets. Amsterdam: North-Holland Publishing Co.

16.

Bryant

(1992) Last interglacial and Holocene trends in sea-level maxima around Australia: Implications for modern rates. Marine Geology 108(2): 209–217.

17.

Bucher

Saenger

(1991) An inventory of Australian estuaries and enclosed marine waters: An overview of results. Australian Geographical Studies 29(2): 370–381.

18.

Clarke

Warwick

(1994) An Approach to Statistical Analysis and Interpretation. Plymouth: Plymouth Marine Laboratory.

19.

Cooper

JAG

(1994) Sedimentary processes in the river-dominated Mvoti estuary, South Africa. Geomorphology 9(4): 271–300.

20.

Cooper

JAG

(2001) Geomorphological variability among microtidal estuaries from the wave- dominated South African coast. Geomorphology 40(1): 99–122.

21.

Cowley

Whitfield

Bell

KNI

(2001) The surf zone ichthyoplankton adjacent to an intermittently open estuary, with evidence of recruitment during marine overwash events. Estuarine, Coastal and Shelf Science 52(3): 339–348.

22.

Dalrymple

Zaitlin

Boyd

(1992) Estuarine facies models: Conceptual basis and stratigraphic implications: perspective. Journal of Sedimentary Research 62(6): 123–145.

23.

Davis

Jr (1989) Texture, composition and provenance of beach sands, Victoria, Australia. Journal of Coastal Research 5(1): 37–47.

24.

Digby

Saenger

Whelan

. (1999) A physical classification of Australian estuaries. In: National River Health Program: Project E1 ed. Lismore, NSW: Centre for Coastal Management Southern Cross University. LWRRDC Occasional Paper 16/99 (Urban Subprogram Report No.9 1999).

25.

Dye

Barros

(2005) Spatial patterns of macrofaunal assemblages in intermittently closed/open coastal lakes in New South Wales, Australia. Estuarine, Coastal and Shelf Science 64(2): 357–371.

26.

Edgar

Barrett

Graddon

. (2000) The conservation significance of estuaries: A classification of Tasmanian estuaries using ecological, physical and demographic attributes as a case study. Biological Conservation 92(3): 383–397.

27.

Escoffier

(1940) The stability of tidal inlets. Shore and Beach 8: 114–115.

28.

Everitt

Dunn

(1991) Applied Multivariate Data Analysis. London: Edward Arnold.

29.

Gale

Pattiaratchi

Ranasinghe

(2006) Vertical mixing processes in intermittently closed and open lakes and lagoons, and the dissolved oxygen response. Estuarine, Coastal and Shelf Science 69(1): 205–216.

30.

Gao

Collins

(1994) Tidal inlet equilibrium, in relation to cross-sectional area and sediment transport patterns. Estuarine, Coastal and Shelf Science 38(2): 157–172.

31.

Haines

Tomlinson

Thom

(2006) Morphometric assessment of intermittently open/closed coastal lagoons in New South Wales, Australia. Estuarine, Coastal and Shelf Science 67(1): 321–332.

32.

Hanes

Ward

Erikson

(2011) Waves and tides responsible for the intermittent closure of the entrance of a small, sheltered tidal wetland at San Francisco, CA. Continental Shelf Research 31(16): 1682–1687.

33.

Harris

Heap

(2003) Environmental management of clastic coastal depositional environments: Inferences from an Australian geomorphic database. Ocean & Coastal Management 46(5): 457–478.

34.

Harris

Heap

Bryce

. (2002) Classification of Australian clastic coastal depositional environments based upon a quantitative analysis of wave, tidal and river power. Journal of Sedimentary Research 72(6): 858–870.

35.

Hart

(2009) Morphodynamics of non-estuarine rivermouth lagoons on high-energy coasts. Journal of Coastal Research 56(2): 1355–1359.

36.

Heap

Bryce

Ryan

. (2001) Australian estuaries and coastal waterways: A geoscience perspective for improved and integrated resource management. A Report to the National Land & Water Resources Audit. Theme 7: Ecosystem Health. Australian Geological Survey Organisation, Record.

37.

Heap

Bryce

Ryan

(2004) Facies evolution of Holocene estuaries and deltas: A large-sample statistical study from Australia. Sedimentary Geology 168(1): 1–17.

38.

Hill

(1994) Some granite landforms of Wilsons Promontory, Victoria. Victorian Naturalist 111: 184–190.

39.

Hinwood

McLean

Wilson

(2012) Non-linear dynamics and attractors for the entrance state of a tidal estuary. Coastal Engineering 61: 20–26.

40.

Hughes

Heap

(2010) National-scale wave energy resource assessment for Australia. Renewable Energy 35(8): 1783–1791.

41.

Hume

Herdendorf

(1988) A geomorphic classification of estuaries and its application to coastal resource management: A New Zealand example. Ocean and Shoreline Management 11(3): 249–274.

42.

Hume

Herdendorf

(1993) On the use of empirical stability relationships for characterising estuaries. Journal of Coastal Research 9(2): 413–422.

43.

Jacobs

Stein

Longcore

(2011) Classification of California estuaries based on natural closure patterns: Templates for restoration and management. Southern California Coastal Water Research Project Technical Publication (619a).

44.

James

Cowley

Whitfield

(2007) Abundance, recruitment and residency of two sparids in an intermittently open estuary in South Africa. African Journal of Marine Science 29(3): 527–538.

45.

Jarrett

(1976) Tidal prism-inlet area relationships (No. WES-GITI-3). Vicksburg, Mississippi: Army Engineer Waterways Experiment Station.

46.

Kench

(1999) Geomorphology of Australian estuaries: Review and prospect. Australian Journal of Ecology 24(4): 367–380.

47.

Kennedy

(2011) Tectonic and geomorphic evolution of estuaries and coasts. In: Wolanski

McLusky

(eds) Treatise on Estuarine and Coastal Science, vol. 1. Waltham: Academic Press, 37–59.

48.

Kirk

Lauder

(2000) Significant coastal lagoon systems in the South Island, New Zealand. Science for Conservation 146: 1173–2946.

49.

Kjerfve

(1994) Coastal lagoons. Elsevier Oceanography Series 60: 1–8.

50.

Kraus

Militello

Todoroff

(2002) Barrier breaching processes and barrier spit breach, Stone Lagoon, California. Shore and Beach 70(4): 21–28

51.

Kraus

Patsch

Munger

(2008) Barrier beach breaching from the lagoon side, with reference to Northern California. Engineer research and development center, Vicksburg, MS Coastal and Hydraulics Lab.

52.

Mondon

Sherwood

Chandler

(2003) Western Victorian estuaries classification project. Report to the Western Coastal Board.Victoria, Australia: Western Coastal Board.

53.

Morris

Turner

(2010) Morphodynamics of intermittently open and closed coastal lagoon entrances: New insights and a conceptual model. Marine Geology 271(1): 55–66.

54.

Mundree

Perissinotto

Nozais

(2003) Seasonal variations in the vertical distribution of benthic microalgae in the upper sediment of the Mdloti Estuary, South Africa. Botanica Marina 46(4): 323–331.

55.

Murray

Radke

Brooke

. (2006) Australia’s near-pristine estuaries: Current knowledge and management. In: Technical Report 63. Australia: CRC Coastal Zone, Estuary and Waterway Management.

56.

Murray-Wallace

Belperio

(1991) The last interglacial shoreline in Australia: A review. Quaternary Science Reviews 10(5): 441–461.

57.

NIWA (2016) Identifying the factors limiting mahinga kai recruitment. Report prepared for Whakaora Te Waihora Partners. Report number CHC2015-078.

58.

Nunes

Adams

(2014) Responses of primary producers to mouth closure in the temporarily open/closed Great Brak Estuary in the warm-temperate region of South Africa. African Journal of Aquatic Science 39(4): 387–394.

59.

O’Brien

(1931) Estuary tidal prisms related to entrance areas. Civil Engineering 1.8: 738–739.

60.

Oliveira

Kjerfve

(1993) Environmental responses of a tropical coastal lagoon system to hydrological variability: Mundau-Manguaba, Brazil. Estuarine, Coastal and Shelf Science 37(6): 575–591.

61.

Peel

McMahon

Finlayson

. (2001) Identification and explanation of continental differences in the variability of annual runoff. Journal of Hydrology 250 (1): 224–240.

62.

Perissinotto

Stretch

Whitfield

. (2010a) Ecosystem functioning of temporarily open/closed estuaries in South Africa. In: Crane

Solomon

(eds) Estuaries: Types, movement patterns and climatical impacts. New York: Nova Science Publishers.

63.

Perissinotto

Stretch

Whitfield

. (2010b) Temporarily Open/ Closed Estuaries in South Africa. New York: Nova Science Publishers.

64.

Piazza

Giuliani

Bellucci

. (2010) PCDD/ Fs in sediments of central Vietnam coastal lagoons: In search of TCDD. Marine Pollution Bulletin 60(12): 2303–2310.

65.

Potter

Chuwen

Hoeksema

. (2010) The concept of an estuary: A definition that incorporates systems which can become closed to the ocean and hypersaline. Estuarine, Coastal and Shelf Science 87(3): 497–500.

66.

Quadros

Rigby

(2010) Construction of a high accuracy seamless, state-wide coastal DEM. Sydney, Australia: FIG Coastal Zone Special Publication.

67.

Ranasinghe

Pattiaratchi

Masselink

(1999) A morphodynamic model to simulate the seasonal closure of tidal inlets. Coastal Engineering 37(1): 1–36.

68.

Ranasinghe

Duong

Uhlenbrook

Roelvink

Stive

(2013) Climate-change impact assessment for inlet-interrupted coastlines. Nature Climate Change 3(1): 83–87.

69.

Riddin

Adams

(2008) Influence of mouth status and water level on the macrophytes in a small temporarily open/closed estuary. Estuarine, Coastal and Shelf Science 79(1): 86–92.

70.

Roy

(1984) New South Wales estuaries- their origin and evolution. In: Thom

(ed.), Developments in Coastal Geomorphology in Australia. San Diego, CA: Academic Press, 99–121.

71.

Roy

Williams

Jones

. (2001) Structure and function of south-east Australian estuaries. Estuarine, Coastal and Shelf Science 53(3): 351–384.

72.

Ruiz-Luna

Berlanga-Robles

(2003) Land use, land cover changes and coastal lagoon surface reduction associated with urban growth in northwest Mexico. Landscape Ecology 18(2): 159–171.

73.

Saintilan

(2004) Relationships between estuarine geomorphology, wetland extent and fish landings in New South Wales estuaries. Estuarine, Coastal and Shelf Science 61(4): 591–601.

74.

Schallenberg

Larned

Hayward

. (2010) Contrasting effects of managed opening regimes on water quality in two intermittently closed and open coastal lakes. Estuarine, Coastal and Shelf Science 86(4): 587–597.

75.

Short

(1996) The role of wave height, period, slope, tide range and embaymentisation in beach classifications: A review. Revista Chilena de Historia Natural 69(4): 589–604.

76.

Short

(2010) Role of geological inheritance in Australian beach morphodynamics. Coastal Engineering 57(2): 92–97.

77.

Slagle

Ryan

WBF

Carbotte

. (2006) Late-stage estuary infilling controlled by limited accommodation space in the Hudson River. Marine Geology 232(3): 181–202.

78.

Terörde

Turpie

(2013) Influence of habitat structure and mouth dynamics on avifauna of intermittently-open estuaries: A study of four small South African estuaries. Estuarine: Coastal and Shelf Science 125: 10–19.

79.

Water Technology (2004) Wave and tidal power assessment for the Victorian coastline. Report to Sustainable Energy Authority No. J121/R01, October 2004.

80.

West

Thorogood

Walford

. (1985) Estuarine Inventory of NSW Fisheries: Bulletin 2. New South Wales, Australia: Department of Agriculture.

81.

Whitfield

(1992) A characterization of southern African estuarine systems. Southern African Journal of Aquatic Sciences 18: 89–103.

82.

Whitfield

Bate

(eds) (2007) A review of information on temporarily open/ closed estuaries in the warm and cool temperate biogeographic regions of South Africa, with particular emphasis on the influence of river flow on these systems. Water Research Commission (WRC) Report No. 1581/1/07.

83.

Woodroffe

Chappell

Thom

. (1989) Depositional model of a macrotidal estuary and floodplain, South Alligator River, Northern Australia. Sedimentology 36(5): 737–756.

84.

Wooldridge

(1994) The effect of periodic inlet closure on recruitment in the estuarine mudprawn, Upogebia africana. In: Dyer

Orth

(eds) Changes in Fluxes in Estuaries: Implications From Science to Management. Fredensborg: Olsen & Olsen, 329–333.