The literature on temperature models for Zostera species growth rate production is often uncertain for theoretical relationships with water temperature (Collier et al., 2017, see their Figure 3). Hence, to minimise complexity, the model-selection criteria were based on published studies of the effects of water temperature on growth rate. The uncertainty of the resulting curvilinear model was verified by assessing the R² and error bar values of the finally derived relationships. It is well known the optimum temperature for seagrass growth is around the peak of a mounded curve (see Figure 1e in Fong et al., 1997), so a theoretical curvilinear relationship of net growth rate with temperature was derived by the leaf marking method used by Holland (1982), which is supported by observations for Zostera marina (see Figure 2 in Moore et al. 1996). In addition, the gross growth rate was found to be linearly related to temperature by Kaldy (2014), so that relationship was derived using net growth rate data from Holland (1982). The Zostera leaf net growth rate, G_net, with water temperature, T°C, was investigated by the study by Scalpone et al. (2020), which confirmed using a curvilinear quadratic relationship, supported by their finding that water temperature limits seagrass growth with an optimal temperature, given by a balance of production with respiration. A curvilinear relationship of growth rate with water temperature was also confirmed by Dennison and Alberte (1985), which gives the mathematical solution by gross growth rate, G_gross, minus the respiration rate R, that is the net growth rate, G_net = G_gross – R, in units mg gdw⁻¹ day⁻¹. The respiration rate has an exponential increase with temperature of the form R = a × e^bT, were “a” and “b” are constants for the seagrass species being modelled (see Figure 3 in Short, 1980), which was confirmed by Plus et al. (2003: Figure 4 by R = 19.14 × e^{(0.11 × T)}), where R is the respiration rate and T is the water temperature. Furthermore, Scalpone et al. (2020) indicated the optimum temperature, T_opt, occurs at the temperature where the respiration rate equals the net growth rate, G_net. Hence, G_net is expected to decrease at temperatures higher than T_opt because R continues to exponentially increase at water temperatures higher than T_opt, thereby giving a curvilinear relationship. That was confirmed by the global review by Nguyen et al. (2021, their Figure 2), which shows seagrass growth rates decrease with increasing respiration rates for twenty-four seagrass species in their Figure 4, including Zostera capricorni.

The increase in Zostera net growth rate production with light intensity is asymptotic to the maximum growth rate, but photo-inhibition decreases photosynthesis at high light intensities in Zostera marina (see Short, 1980: Figure 4). More recently, Beer and Björk (2000) found Halodule wrightii had gross photosynthesis curvilinear with the irradiance light intensity, while Campbell et al. (2008, see their Figure 3) showed a curvilinear relationship of photosynthesis with light intensity due to photo-inhibition to begin at about 1200 μmol photon m⁻² sec⁻¹ for three seagrass Halophila spp. That was confirmed by Zhang et al. (2022, their Figure 2), who found a curvilinear relationship of Zostera marina productivity with light intensity duration. Similarly, photo-inhibition measured at 0.5 m depth inside the Zostera capricorni bed by Holland (1982) in The Tuggerah Lakes, caused a curvilinear relationship of growth rate with light intensity. On the basis of the literature, a curvilinear relationship of growth rate with light intensity was statistically tested using the data from Holland (1982).

The seagrass growth rate with sediment phosphorus concentrations by Fong et al. (1997, see their Figure 1c) increases approximately linearly up to a maximum and then levels out. However, Harris (1977) only measured the Zostera capricorni biomass and found a linear relationship with sediment total phosphorus concentrations in the Lake Illawarra coastal lagoon. As the data in Holland (1982) and from Harris (1977) are from nearby lagoons, the Lake Illawarra data and relationship were assumed to apply to Lake Munmorah in the Tuggerah Lakes used by Holland (1982).

The literature shows that the derived equations based on the results in Holland (1982) and Harris (1977) are valid and could be applied to other seagrass species in estuaries and coastal lagoons. The above literature shows that many seagrass species have curvilinear relationships. For example, the study by Collier et al. (2017, their Figure 4) showed Zostera muelleri (subspecies of Z. capricorni), Cymodocea Serrulata and Halodule uninervis had a curvilinear relationship with net productivity (equivalent to net growth rate in Holland (1982). In addition, Campbell et al. (2008) showed the seagrass species of Halophila spinulosa, Halophila decipiens, Halophila ovalis, Cymodocea serrulata, and Syringodium isoetifolium had curvilinear relationships of photosynthesis with light intensity. That suggests related species in other estuaries and saline coastal lagoons may have similar relationships to 2^nd Order curvilinear obtained using data in Holland (1982). For example, the online literature shows Cymodocea serrulata is in the family Cymodoceaceae and has 16 seagrass species in tropical and subtropical coastal waters. The common seagrasses are Syringodium isoetifolium, Cymodocea nodosa, Halodule uninervis, and Cymodocea rotundata, and in the tropical and subtropical Indo-Pacific is Syringodium isoetifolium. Hence, the curvilinear relationships of these seagrass species in estuaries and saline coastal lagoons is worthy of further research. The literatures supporting these equations are shown in the methods and the references for key findings are shown in the Discussion.

The light and water temperature are shown in Attachment Al (note A1 was altered to Al i.e., ‘lower case L’) and is described in Holland (1982): Chapter 2 and Zostera growth rates in Chapter 4 undertaken at an inshore site in 1.0 m water depth called Munmorah Flats, MF, in the north-west of Lake Munmorah (Holland, 1982: Figure 6.1). The site was selected with a low catchment population, so the effects of water temperature and light conditions showed natural changes in healthy areas of Zostera without effects of the moderate nutrient inputs. Eleven net growth rate measurements in response to temperature and light conditions were undertaken over a 17-month period from October 1978 to February 1980. The incident light intensities, I₀, at 0.5 m above the water surface and the light intensity at 0.5 m inside the Zostera bed, I_d, were measured by Holland (1982) with a Li-Cor quantum sensor measuring photosynthetically active radiation (PAR) units of radiation (µmol m⁻² sec⁻¹). Thermal effects on seagrass growth rates were measured in the laboratory at low light by Holland (1982, Table 6.6 “Growth rates of Zostera shoots after a range of exposure times”) by measuring the thermal reduction in growth rates at fixed water temperatures over time (see Section 3.5.1).

Holland (1982) used the leaf marking method in the field, which assesses growth directly and only requires measurement of the leaf loss rate, L (Verhagen & Nienhuis, 1983), rather than the respiration rate required by the oxygen evolution method. The leaf growth rate was measured by the increase in leaf length biomass using the “Increase in Plant Tissue method”, called here net growth rate, G_net. Note that the net growth rates have units of mg gdw⁻¹ day⁻¹, called “comparative growth rates” by Holland (1982), which are equivalent to the nomenclature for seagrass production in the literature. The Zostera biomass was calculated by the monthly average shoot weight multiplied by the shoot frequency in Holland (1982), Chapter 5, Table 5.2.

Leaf loss rates are an important part of seagrass processes. It was noted by Moore et al. (1996) that leaf loss rates were highest in summer, indicating the loss rates are related to temperature, so the relationship of loss rates with water temperature was derived using the Zostera percentage viability data in Holland (1982). To derive that relationship, loss rates were estimated using the percentage viability of the Zostera leaves, measured in relation to the net rates, from Appendix A4 “Percentage Viability”. The Loss rates were estimated by ([1 - % 1%-viability/100] x Table 6.1 “net growth rate”) mg gdw⁻¹ day⁻¹, shown in Table II. The loss rates allowed estimation of the gross growth rates, G_gross, by G_gross = G_net (from Table 6.1) + L, and the estimated gross growth rates are also shown in Table II. Note that decreases in viability are due to damage and bacterial degradation of detritus (Patrıćio & Marques, 2006).

Statistical analyses were undertaken by Microsoft Excel, least squares regression using curvilinear polynomial types, including 2^nd Order quadratic for one independent variable, such as water temperature and subsurface light intensity, compared with a given seagrass variable, such as Zostera growth rate, and linear and exponential for other relationships. The equations were evaluated on the basis of overall goodness of fit statistics using R² with p-value < 0.05 as acceptable. The effect of a water temperature increase due to recent climate change effects showed that the temperature increase of 1.4°C to 28°C on seagrass growth rates was consistent with the derived curvilinear water temperature equation using data from Holland (1982). The standard deviation, σ, for each point on the regression curve is shown using Excel Vertical Error Bars. The obtained significant R² values for higher-order curvilinear equations for the seasonal variation of water temperature and light intensity detected abnormal changes compared with the usual seasonal pattern. More complex statistical analyses were not expected to provide additional information for estuary managers to investigate variations in Zostera biomass and growth rate monitoring results.

The data sources from Holland (1982) used for analyses are shown in the data summary Table I, and could be found by clicking on the links below Table I and in the references and using the locations shown in the table. Hence, the actual data is also available by using the internet links shown below the summary table and in the references. To avoid confusion with reference to other data, the Holland (1982) sources are shown in italics with quotation marks.

Data gaps or limitations were investigated by having the data cross-checked for consistency between relationships with temperature and light conditions, and, where possible, with published data from other sources. In that regard, Ives et al. (2003) suggested that water temperature and light intensities be graphed and a curvilinear regression used to indicate significant variations and possible adjustment for consistency with the normal seasonal variation. For example, Plaisted et al. (2022, their Figure 4) showed Zostera marina abundance has a curvilinear relationship with water temperature, and Nielsen et al. (2002) found Zostera marina growth at depth depended on water turbidity. In addition, Jakeman et al. (2006) suggested consistency of the data be tested by an iterative process until convergence was obtained, so the various temperature relations were based on a reliable database. A similar iterative process was used for review of the more variable and difficult to measure light data under field conditions from Holland (1982). Note that in Section 3.2, it was noted that the Zostera biomass was lower at the low catchment population reference site than in the more highly populated areas of Lake Munmorah, indicating differences in sediment total phosphorus. Hence, no autocorrelation statistics, such as the Akaike information criterion (AIC), were required as no model selection was required. That is, it was likely that no spatial autocorrelation was present for the completely different sites. However, when comparing changes over time at the same reference site in Section 3.8, it was assumed that autocorrelation over time allowed estimation of unsampled months related to the measured values by tracking the graph of changes over time. The adjusted values from these processes are shown in Table II. Estimated values for dates unable to be sampled are also shown, including estimated light intensities in the Zostera bed, loss rates, and gross growth rates, along with estimated biomass for months not sampled. Importantly, reference seagrass locations need to be used to provide a comparison with the selected routine sampling site. For example, broad-scale seagrass mapping by King and Holland (1986, see their Table I) for Zostera biomass in the Tuggerah Lakes coastal lagoon was undertaken using visually categorized percentage cover. The broad-scale sampling included the Munmorah Powered Station cooling water area in Budgewoi Lake, compared with reference areas Munmorah Lake and Tuggerah Lake. Recently, reference locations have been used to standardize the estimated seagrass biomass and areas during satellite image mapping in Southern Thailand (Koedsin et al., 2016).

The data by Holland (1982) covered a range of temperatures from the minimum of 13.8°C in July 1979, to 26.6°C in March 1978, and a range of incident light intensities of 979 in July to 2500 μmol m⁻² sec⁻¹ in January 1980 (Table II). The range of Zostera net rates varied from 4.8 in July to 13.8 mg gdw⁻¹ day⁻¹ in September, 1979, with a 10-fold range in biomass from 2.6 in August, 1979, to 26.9 g dw m⁻² in October, 1978. To test the consistency of the actual measured net growth rates (see Table 6.1) with temperature (Appendix Al, page 194), due to potential environmental changes during the 17 months of sampling were considered. A quadratic curvilinear 2^nd Order relationship was derived for net growth rates with water temperature, as well as for Light intensities at 0.5 m inside the Zostera capricorni bed at the MF sampling site, using the data from Holland (1982), shown with some adjustments in the list in Table II. The water temperature relationship revealed the following net rate outliers relative to the others, which were investigated, and adjusted if necessary: a higher than expected 13.8 at 18.1°C in September, 1979, and lower than expected 10.7 mg gdw⁻¹ day^–1 in October, 1979, at the warmer 20.4°C. That also indicated the continuing decrease in net rates in November and December 1979 to 9.9 and 9.2 mg gdw⁻¹ day^–1, while the temperatures increased to 24.4 and 24.3, respectively.

In the wider area of Lake Munmorah, where about half the nearshore catchment was highly urbanized, the Zostera leaf biomass at about 28 g dw m⁻² (King & Hodgson, 1986 their Tables 3 and 4), was about double that at the MF site of about 12 g dw m⁻², indicating a higher nutrient input to the main body of the lake. As de Boer (2007) indicated, moderate levels of sediment nutrients could promote seagrass growth; the total phosphorus (TP) content of seagrass bed sediments and its relationship with Zostera biomass were derived from the study by Harris (1977). Consequently, the stability of the Zostera leaf biomass at MF and in Lake Munmorah was compared in Section 3.8, and it was shown that the higher TP input maintained a higher and more stable biomass. The relative stability of the biomass indicated the importance of the three main environmental characteristics, and the detection of significant biomass changes over time could be used by estuary managers as part of their controls on catchment sediment and nutrient inputs.

The list of data adjustments and estimation of missing data is shown here. The a), b), c) etc. refer to the superscripts in the table headings.

a) Water temperatures in October 1978 and January 1979 from Holland (1982). Figure 2.4. Table 6.1 net rates had adjustments: September 1979 to 12.4 by 13.8 - 1.4 loss rate from 1.4 = (13.8 × [1 - 90 viability/100]) mg gdw⁻¹ day^–1 by viability of 90% estimated from Figure 6.2 for consistency with other net rates. October, 1979 10.7 to 13.2 = (10.7 + loss rate 2.5), see Section 3.2. November and December 1979 heatwave effects by net rate = gross rate - loss rate by (4) in Section 3.4, but the measured rates of 9.9 and 9.2 were consistent with the measured biomass in Table 5.2. November 1978 net rate at the highest measured water temperature of 26.6°C and potential maximum average daytime monthly average of 28°C, with Maximum T due to climate change 1.4°C added to 26.6°C, both estimated using (4). Not sampled and consistency with measured and adjusted estimated by G_gross – L.

b) Loss by net rate Table 5.1 × (1 - %viability/100) mg gdw⁻¹ day⁻¹ in Appendix A4. Loss rates adjusted consistently with Table 6.1 by: December 1978 adjusted from 5.1 to 4.1 by (3) November, 1979 5.9 to 6.6. July 1979 viability was shown as 90% in Figure 6.2 (down from 100% in Appendix A4), and the loss rate of 0.48 (4.8 × [1 - 90/100]) was consistent with a temperature of 13.8°C in Fig. 4. (3) indicated the loss rate in January, 1980 was about 7.2 at a temperature of 24.7°C rather than 4.5 (net rate 12.9 × [1 - 65%/100]), requiring a viability of 44%, giving 7.2 (12.9 × [1 - 44/100]) mg gdw-1 day–1. The not-sampled and November 1978, and maximum water temperature, 28°C, loss rates were estimated by (3).

c) Gross Growth Rates were estimated by measuring net rates in Table 6.1 plus L and not sampled by (2) derived from Fig. 3.

d) Incident light intensities, I0, are shown in Table 2.2 and the values affected by cloud or rain in March, June and August, 1979, and the unable to sample October, 1978 and January, 1979, were estimated by proportioning with the Australian Bureau of Meteorology (BoM) average monthly light intensities at the nearby Norah Head Lighthouse. In addition, the October 1979 value of 1320 was low compared to other intensities, so it was adjusted to 1468 by (1320 µmol m⁻² sec⁻¹ × October MJ m-2 17.315/September 15.57). The adjusted values were consistent with Fig. 9 and derived (9) light intensities inside the Zostera bed with incident light intensities.

e) The not sampled % Light intensity reaching 0.5m inside the Zostera bed was back calculated from the 100 x Id/I0, where Id is the estimated bed light intensity and I0 is the incident intensity. In February 1979, the % was adjusted from 86.7 to 66.1.

f) Light intensities reaching 0.5m inside the Zostera bed, Id, were estimated by Id = I0 × % light penetration/100, using incident light intensities in Table 2.2 and % light penetration in Appendix Al, page 196. Not sampled ID estimated by (9).

g) Measured biomass is in Holland (1982) Table 5.2. The not sampled Zostera biomass, g dw m⁻², values were estimated by (11): B = 4.0106 x G_net − 32.885 relationship of biomass in Table 5.2 with the measured net growth rates in Table 6.1. October 1978 biomass of 26.9 was adjusted for consistency with the measured net rate of 12.8 mg gdw⁻¹ day^–1 to 18.5 gdw m⁻² by (11). The July 1979 biomass was estimated by proportioning the August biomass, 2.6 g dw m^−2, by the July and August net rates, giving 1.3 g dw m⁻² (2.6 × 4.8/9.4).

The seasonal water temperature variation derived from raw data in Figure 2.4 in Holland (1982) is shown in Fig. 1. Although seasonal water temperature variations are usually assumed to vary as a sine curve (Saborowski & Hünerlage, 2022), it is unlikely the summer and winter conditions are the same each year, so the temperature variations were assumed to be describe by a curvilinear relationship.

The curvilinear relationship in Fig. 1 was obtained using the 4th-order polynomial:

(1) Water temperature = − 0.0073 × ( m o n t h s i n c e O c t o b e r , 1978 ) 4 + 0.2778 × ( m o n t h s i n c e = O c t o b e r , 1978 ) 3 − 3.4095 × ( m o n t h s i n c e O c t o b e r , 1978 ) 2 + 14.289 × ( m o n t h s i n c e O c t o b e r , 1978 ) + 7.9803

with R² = 0.8634, n = 17, p > 0.001. Note (month since October 1978)ⁿ is calculated by POWER (number, n). The low temperature in December 1978 was investigated by examining the relationship between gross, loss, and net rates with temperature in Figs. 3 and 4, which showed the measured temperature was consistent, indicating an abnormally low temperature at the time of measuring the net growth rate of 12.8 mg gdw⁻¹ day^–1.

Due to environmental variations and difficult field measurements, the net growth rates showed variations in September, October, November, and December 1979 from the theoretical quadratic curve with water temperature. The September 1979 net rate of 13.8 was adjusted to 12.4 by subtracting the loss rate estimated as 1.4 (13.8 × [1 − 90/100]) mg⁻ gdw⁻¹ day⁻¹ by an estimated viability of 90%, giving a consistent loss rate with water temperature in Fig. 4. In October, 1979, the loss rate of 2.5 (10.7 × [1 − viability 77/100]) in Appendix A4: “Percentage Viability Lake Budgewoi Test Sites” was consistent with the water temperature/loss rate relationship in Fig. 4, but the net rate of 10.7 appeared low for a temperature of 20.4°C, suggesting a net rate of 13.2 (10.7 + 2.5) mg gdw⁻¹ day^–1. That was tested by the relationship of gross growth rate with water temperature in Fig. 3. The probable gross rate of 15.7 (net 13.2 + 2.5 loss rate) mg gdw⁻¹ day^–1plotted against water temperature was consistent with the regression obtained by (2). The adjusted net rates are shown in Table II.

As the growth rates during the southern spring/summer of November and December 1979 of 9.9 and 9.2 mg gdw⁻¹ day⁻¹ were much lower than over the same period in 1978, the cause was investigated. It was noted that a heatwave event occurred on those dates, causing the Zostera bed water temperatures to reach 28.7°C and 29.5°C, respectively (page 24 in Holland, 1982), probably suppressing the growth rate measurements. The reported net rates were measured at lower temperatures of 24.4°C and 24.3 °C, apparently after the heatwave had finished. However, it appears the plants had been damaged in some way by the heatwave, so the expected rates were estimated from the relationship of gross rates with temperature; see Fig. 3, minus the estimated and adjusted loss rates in Table II, giving the adjusted net rates of 13.1 and 13.0 mg gdw⁻¹ day^–1, respectively, which are also shown in Table II. The seasonal variation in net growth rate is shown in Fig. 2 from Figure 6.3: in Holland (1982), with adjustments for effects of water temperature variations, potential light intensity effects, and loss rates.

The study by Santamaŕia and van Vierssen (1997) showed that water temperature governed the Zostera biological processes, and the more variable light conditions modified the response to water temperature, so water temperature was examined first. As the gross growth rates are estimated from G_net + L in Table II, the gross growth rates are compared with water temperatures from Appendix Al (Holland, 1982) in Fig. 3.

The relationship obtained in Fig. 3 was:

(2) G r o s s g r o w t h r a t e ( m g g d w − 1 day − 1 ) = 1.148 × W a t e r t e m p e r a t u r e ( ∘ C ) − 8.3893

R² = 0.9573, n = 11, p > 0.001. To complete the representation, the loss rate relationship with water temperature is shown in Fig. 4.

The relationship obtained in Fig. 4 is given by:

(3) L o s s R a t e ( m g g d w − 1 day − − 1 ) = 0.0143 × e ( 0.2522 × T )

with R² = 0.9889, n = 9, p > 0.001. As the net rate could be estimated by the difference between gross growth rates by (2), minus the loss rate, L, by (3), the net growth rate could be estimated by:

(4) N e t g r o w t h r a t e = 1.148 × T − 8.3893 − 0.0143 × e ( 0.2522 × T )

Due to difficulties with field measurements, two of the 11 incident light intensity measurements, corresponding with G_net in October 1978 and August 1979, could not able to be measured, so they were estimated using the local BoM measurements at the Norah Head Lighthouse in 2022, with details of those estimates shown in the footnotes to Table II.

The biomass of Zostera was estimated by Holland (1982) in Table 5. “Zostera reproductive shoots shoot weight, frequency and biomass.” by multiplying the mean shoot weight (mg dw shoot⁻¹) by the shoot frequency (g dw m⁻²). The Zostera biomass was assumed to be related to the net growth rate, so it was investigated in Fig. 11.

The relationship from Fig. 11 is given by:

(11) Z o s t e r a c a p r i c o r n i b i o m a s s ( g d w m − 2 ) = 4.0106 × N e t g r o w t h r a t e ( m g d w g d w − 1 day − 1 ) − 32.885

with R² = 0.8976, n = 6, p > 0.01. The regression indicated Zostera biomass to become nil at G_net of 8.2 mg gdw⁻¹ day⁻¹, which was likely to occur during the southern winter conditions of low light and temperatures, measured in July 1979 at 979 μmol m⁻² sec⁻¹ and the temperature of 13.8°C. However, the measured biomass in August 1979 was unexpectedly low at 2.6 g dw m⁻², which appears due to the likely low nutrient catchment inputs to the sampling site. The maximum G_net of 13.4 mg g dw⁻¹ day⁻¹, see Fig. 10, was estimated to produce a biomass of 20.9 g dw m⁻² using (11). That was investigated by the seasonal trend at the MF sampling site, compared to the long−term biomass changes in the whole of Lake Munmorah using Stability Charts.

The stability of a freshwater aquatic ecosystem was investigated by Ives et al. (2003), who found the system was stable when the variability of the long-term abundance was less than the variability of the environmental factors such as temperature, nutrient inputs, or other factors causing the system’s abundance to change. Hence, Stability Charts for the seasonal variation of Zostera biomass with upper and lower standard deviations from the long-term average (derived from the study by Carr et al., 2012) are used here to show Zostera biomass changes. The stability Chart helps to identify responses to temperature and light conditions, along with catchment nutrient inputs. As the relationships obtained are reasonably consistent, the seasonal changes in Zostera biomass at the MF site from Table 5.2 in Holland (1982), including the estimated not-sampled months in Table II using (11), were tracked by a graph of changes over time. The missing biomass values in Fig. 12 were estimated from the regression in Fig. 11, which has a standard deviation for differences between each reported and estimated biomass value of σ ± 1.65, giving confidence about 13% of the average biomass of 12.3 g dw m⁻². That gave a realistic estimate of the missing biomass.

The Stability Chart shows changes that may require further investigation of potentially unknown environmental effects. For example, Hodgson and Bucher (2023) showed the long-term changes for average total seagrass area ± 1 standard deviation, σ, in coastal lagoons gave a reasonable estimate of the variability. They found in Lake Macquarie, 1σ accounted for about 85% of the long-term variability. Hence, it was assumed the seasonal Zostera biomass changes at MF could also be compared to the average ± 1σ, which is shown in Fig. 12 Stability Chart for biomass changes.

The Stability Chart accounted for 70.6% of the observations with a wide range about the average of 12.3 from 2.6 to 21.2 g dw m⁻² around the average. In addition, the above estimated biomass of 20.9 g dw m⁻² for maximum G_net 13.4 mg gdw⁻¹ day⁻¹ was almost the same as the upper biomass range, indicating consistency of the light intensity-derived net rate with biomass. However, the chart shows biomass was related to the heatwave-affected net rates of 9.9 and 9.2 in November and December 1979, whereas Fig. 5 shows the higher adjusted net rates of 13.1 and 13.0 mg gdw⁻¹ day⁻¹ in Table II were related to water temperature. That finding is supported by those of George et al. (2018) that high water temperatures have a greater effect on biomass than photosynthesis. Fig. 12 also shows the southern autumn and winter, May to August, 1979, levels at or below the lower investigation level. The cause was investigated by examining the August 1980 seagrass map (see Figure 4 in King & Holland, 1986), which shows that nearly all the MF sampling sites used by Holland (1982) were covered with a high abundance of Ruppia and no Zostera in the area covered by the Ruppia. Catchment nutrient inputs were observed by Cho et al. (2009), causing an increased abundance of Ruppia maritima, the same species that occurs in the Tuggerah Lakes (King & Holland, 1986). However, Holland (1982) did not find excessive Ruppia or macroalgae at the MF site. That indicates a change in environmental conditions in August 1980, so the change was investigated by a Stability Chart of the long-term changes in Zostera biomass for the whole of Lake Munmorah by King and Hodgson (1986), which is highly urbanized in the southern half, shown here in Fig. 13.

The Fig. 13 shows that the whole lake chart accounted for 82% of the variability, with a range about the average of 28.1 from 16 to 40.8 g dw m⁻². The lowest biomass of 16.0 in August 1980, was about 13-fold higher than the 2.6 g dw m⁻² in August 1979, at the MF site, while the whole lake average was about twice that at the MF site, with the highest at 40.8 g dw m⁻² during southern winter, June 1981. Those observations indicate that the higher biomass and lower variability than at MF could be due to nutrient inputs from the more densely urbanized nearshore catchment in the southern areas of Lake Munmorah. Consequently, the estimated MF site biomass of 20.9 g dw m⁻²; for maximum G_net of 13.4 mg^−gdw−day was now near the lower biomass range for Lake Munmorah. That suggests, for the maximum G_net to be related to the upper biomass in Lake Munmorah of 36.7 g dw m⁻², the whole of Fig. 10 quadratic relationship may have to be increased to give a higher maximum G_net in the lake of about 17.4 (36.7 + 32.885/4.0106) mg gdw⁻¹ day⁻¹ by (11). Hence, the potential G_net increase by sediment TP concentrations was tested by examining the relationship of Zostera biomass with sediment total phosphorus.

By assuming the Zostera biomass and sediment total phosphorus in Lake Illawarra (the location of the lagoon is shown in Figure 1.2, page 11 in Harris (1977), 34°32′32.2″S, 150°49′32.8″E) gives an approximation to Lake Munmorah biomass, the data from Harris (1977), in the summary Table I, was used to develop the relationship shown in Fig. 14.

The relationship obtained was:

(12) Z ⁢ o ⁢ s ⁢ t ⁢ e ⁢ r ⁢ a b ⁢ i ⁢ o ⁢ m ⁢ a ⁢ s ⁢ s ⁢ ( g ⁡ d ⁢ w ⁢ m − 2 ) = 2.031 × S ⁢ e ⁢ d ⁢ i ⁢ m ⁢ e ⁢ n ⁢ t ⁢ t ⁢ o t a l p h ⁡ o s p h ⁡ o r u s ( μ g gdw − 1 ) − 112.83

R² = 0.8469, n = 18, p > 0.001.

The Fig. 14 shows that a sediment TP < about 55 μg gdw⁻¹ may not have influenced water temperature and light intensity quadratic relationships with G_net at the MF sampling site selected by Holland (1982) to provide relatively natural conditions, and likely provided baseline conditions for Zostera growth rates. Hence, the probable low TP may explain the higher variability, without heatwave effects in November and December, of the biomass at MF in Fig. 12, compared to the more stable and higher biomass, with two southern winter and 5 summer surveys, in the urbanized areas of Lake Munmorah in Fig. 13.

The likely higher TP input to Lake Munmorah was estimated from the average biomass of 28.1 g dw m⁻² in Fig. 13 by (12) as about 69 ([28.1 + 112.83]/2.031) μg gdw⁻¹, a moderate TP at the low end of the range in Fig. 13. As the maximum G_net at MF 13.4 was related to the upper biomass, the same may have occurred in Lake Munmorah, so the TP required to increase maximum G_net to above 17.4 mg gdw⁻¹ day⁻¹ for the upper biomass was estimated as 74 ([36.7 +112.83]/2.031) μg gdw⁻¹. Although the maximum G_net was about 30% higher than at MF, the TP was only about 12% higher, and gave a significant increase in average biomass to 28.1 compared to 12.3 g dw m⁻² at MF with improved stability. That is an important finding and suggests, with further research, the possibility of maintaining Zostera biomass in estuaries and coastal lagoons by control of important catchment nutrient inputs, such as total phosphorus.