nvestigating the influences of tidal inundation and surface elevation on the establishment and early development of mangroves for application in understanding mangrove rehabilitation techniques 1 4

Although only a subset of natural mangrove surface gradients and floral species can be found at the reference forests, these sites nonetheless present the best reference as more than 90%

Chapter 3 – Surface elevation is an important factor in

achieving mangrove rehabilitation success

3.1 Introduction

The number of mangrove rehabilitation programmes implemented world-wide is extensive, the majority of which have been executed by governments and local stakeholders to restore forest cover and habitat functionality of degraded mangrove systems (Katon et al., 2000; Barbier, 2006) This commonly takes the form of planting and replanting (when previous attempts have failed) of mangrove propagules and/or seedlings, with low survival rates (Samson & Rollon, 2008) Yet, rehabilitation programmes can be successful if rehabilitation methods involve matching environmental conditions to the autecology of mangrove species

This can be achieved if rehabilitation methods prioritise and optimise structural manipulation to remove barriers (e.g dike walls) and allow for the regeneration of mangroves via natural recruitment and establishment This encompasses hydrological and substrate regrading for re-establishment of appropriate hydrologic regime and elevation ranges (Stanley & Lewis, 2011) since hydrology and surface elevations are major leverage points influencing mangrove establishment, survival and development because of its control on inundation regimes Notably, inundation is one of the primary factors in determining the establishment, survival and growth of mangroves because of its influence on secondary factors such as oxygen availability, salinity and pH (Krauss

et al., 2008)

Hence, knowledge of appropriate surface elevation and its influence on inundation frequency, period and depth may be one of the more important factors that determine

the overall success of mangrove rehabilitation projects (Lewis, 2005; Gilman & Ellison, 2007) Through executing a rehabilitation project, this study investigated the contributions of two aspects of hydrologic restoration to the successful colonisation of mangrove vegetation in rehabilitated sites The study focused on (i) restoring a natural tidal inundation regime via strategic breaching and (ii) regrading selected areas to suitable surface elevations

3.2 Materials and Methods

3.2.1 Study area

The Western coastline in South Sulawesi Province, Indonesia, experiences a monsoonal climate The northwest monsoon (December – March) is characterised by high precipitation and strong winds while the southeast monsoon (June – September) brings negligible rainfall (Visser et al., 2004) This coastline experiences a tidal range

of 1.6 m with a mean tidal range of 0.95 m, with Mean High Water Spring at 1.33 m Chart Datum

This study was conducted in three locations near the fishing villages of Kurri Caddi and Kurri Lompo (5° 01' 57" S, 119° 28' 04" E, Figure 3.1) The main study site consists of 29 disused aquaculture ponds, covering 21.5 ha Before rehabilitation, the ponds were used for semi-intensive farming of brackish water shrimp and milkfish polyculture These ponds were created in the early 1980s and originally operated by a Korean venture They were then bought and managed by the University of Muhammadiyah, Makassar (UNISMUH) In the past, neighbouring communities had

no access or use rights in the ponds However, after an agreement was reached between Mangrove Action Project – Indonesia (MAP-I) (now Blue Forests Indonesia),

aquaculture ponds The community now holds rights to access non-timber resources in some ponds and collaborate with MAP-I in participatory action research involving mangroves, aquaculture ponds and their own rice-fields Interested members of the community are also involved in a multi-stakeholder working group involving academic, governmental and NGO partners, which serve as to manage and advise mangrove rehabilitation

Riverine/lower estuarine mangroves surround these aquaculture ponds Two reference forests, comprised solely of coastal and riverine greenbelts, located at a distance of 2.3

km and 0.05 km from the aquaculture ponds and approximately 200 and 50 m wide respectively, were surveyed (Figure 3.1b) Although only a subset of natural mangrove surface gradients and floral species can be found at the reference forests, these sites nonetheless present the best reference as more than 90% of mangroves in Maros and neighbouring Pangkep District have been converted into aquaculture

Figure 3.1: (a) Regional setting of Makassar (black box), South Sulawesi, Indonesia; (b) former aquaculture ponds at Kuri Caddi and reference forests at both Kuri Caddi and Kuri Lompo; and (c) broken lines delineate the disused aquaculture ponds in Kuri Caddi, extracted from Google Earth (dated February 2014)

3.2.2 Field data collection

Pre-rehabilitation mapping of abandoned aquaculture ponds – In September 2013, a

Trimble Real Time Kinematic GPS (RTK-GPS) was used to establish nine elevation benchmarks throughout the site in the WGS84 coordinate system From these benchmarks, a topographic survey of the former aquaculture ponds was conducted

sampling point resulted from the mean of three readings taken As the topography of the ponds was not highly complex, surface elevation measurements were sampled along both dike walls and in the ponds at (approximately) every 3 metres Sampling point density was increased when abrupt changes in surface elevation were observed (i.e sudden drop between the dike walls into aquaculture pond)

Rehabilitation of aquaculture ponds – In November 2013, rehabilitation works were

implemented in selected areas in the ponds through the use of an excavator and community labour with hand tools Dike walls were strategically breached, based on the elevation gradients observed from the previous topographic mapping exercise, in order to restore hydrological flows Some were regraded entirely to produce substrate

at an appropriate surface elevation for mangroves, relative to sea level (Figure 3.2) A pile of broken branches was also deployed in one pond, designed to trap floating propagules Across the 29 aquaculture ponds, a total of 16 breaches were made in the dike walls Since seed banks are generally absent in mangroves (Harun-or-Rashid et al., 2009), hand broadcasting of locally-collected propagules was conducted at high tide to overcome propagule availability as a limiting factor Approximately 206250

Aegiceras corniculatum (55 kg; 3750 propagules kg-1), 1500 Avicennia sp (10 kg; 150

propagules kg-1), 2380 Bruguiera gymnorrhiza (70 kg; 34 propagules kg-1) and 8600

Ceriops tagal (47 kg; 183 propagules kg-1) were broadcasted in December 2013

Rhizophora spp., Sonneratia spp and B cylindrica propagules were also broadcasted,

but in unknown quantities To facilitate seedling establishment, the ponds were subsequently left untouched

Figure 3.2: (a) Dike walls that have undergone strategic breaching, and (b) regrading of selective dike walls to produce substrate at lower surface elevations (foreground) Red arrows point to existing dike walls

3.2.3 Post-rehabilitation vegetation survey in aquaculture ponds and reference mangrove forests

A second topographic survey using a Total Station anchored to the existing benchmarks was conducted in June 2014 First, an elevation survey was conducted to measure changes in surface elevation of regraded areas Then, a second survey was conducted to measure the surface elevation at which mangrove vegetation had established inside the newly restored site Vegetation was identified to species level

and categorised into three size classes – seedling (< 0.7 m in height), sapling (> 0.7 m

in height but < 7 cm Diameter-Breast-Height; DBH) and tree (≥ 7 cm DBH)

Separately, this survey was repeated in the reference forests (i.e the two natural mangrove forests near Kuri Caddi and Kuri Lompo) in order to produce the elevation envelope (inter-quartile range) of the natural mangrove elevation range for different species

3.2.4 Genera-specific surface elevation envelopes and prediction maps

ArcGIS was used to generate a Digital Elevation Map (DEM) of the ponds using the Linear interpolation algorithm with the chosen grid cell size (resolution) of 1

The statistical computing software R 3.1.2 (R development core team, 2014) was used

to define species-specific elevation envelopes This was computed separately for trees and seedlings, and if they were established in reference forests or aquaculture ponds Thereafter, species-specific elevation envelopes for trees in reference forests were checked for normality and equality of variances before conducting a t-test Given that the species-specific elevation envelope between two species (per genus) were statistically similar (p-value < 0.05), they were combined to give a genera-specific

elevation envelope for genera Avicennia, Rhizophora and Sonneratia

The Raster Calculator in ArcGIS was used to delineate areas in the aquaculture ponds exhibiting the exact elevation range in each genera-specific elevation envelope data This created a prediction map of aquaculture ponds of predicted future mangrove

growth for each of the three genera – Avicennia, Rhizophora and Sonneratia

3.3 Results

3.3.1 Vegetation established in aquaculture ponds and reference mangrove forests

A total of 471 seedlings and saplings and 180 trees were surveyed in aquaculture ponds compared to the 213 seedlings and saplings and 140 trees in reference forests

(Table 3.1) In the ponds, Rhizophora was the most abundant seedling/sapling genera (n = 254) with Avicennia as the second most abundant seedling/sapling genera (n = 97) The top three most abundant seedling/sapling species were R mucronata (n = 116), R styolsa (n = 84) and R apiculata (n = 54) For trees in aquaculture ponds, the opposite is observed wherein Avicennia was the most abundant genera (n = 55), followed by Rhizophora (n = 52) The top three most abundant tree species are R

mucronata (n = 45), L racemosa (n = 38) and A marina (n = 30) In reference forests,

similarly, Rhizophora was the most abundant seedling/sapling genera (n = 138) with

Avicennia as the second most abundant seedling/sapling genera (n = 50) The top three

most abundant seedling/sapling species were R mucronata (n = 72), R apiculata (n = 58) and A marina (n = 30) For trees in reference forests, similarly, Avicennia was the most abundant genera (n = 59), followed by Rhizophora (n = 40) The top three most abundant trees species were A marina (n = 38), R mucronata (n = 37), and S alba (n

= 38)

Across both aquaculture ponds and reference forests, Bruguiera spp was observed to

be present in low numbers 16 Bruguiera spp seedlings/saplings were surveyed in the aquaculture ponds Bruguiera spp trees were absent in both reference forests surveyed

(Table 3.1)

Table 3.1: Number of seedlings/saplings and trees surveyed across aquaculture ponds and reference forests

Seedlings/saplings surveyed have established at similar surface elevation ranges in both aquaculture ponds (Figure 3.3a; -1.511 m ≤ x ≤ 0.228 m WGS 84) and reference forests (Figure 3.3b; -1.255 m ≤ x ≤ 0.073 m WGS 84) Of all seedling species

surveyed in aquaculture ponds, the elevation envelope occupied by A rumphiana was the widest, followed by A marina and B gymnorhiza (Figure 3.3a; -1.085 m ≤ x ≤ -

0.218 m WGS 84; -0.994 m ≤ x ≤ -0.269 m WGS 84; -0.861 m ≤ x ≤ -0.219 m WGS

84) In reference forests, the species were R mucronata, R apiculata and Sonneratia

alba (Figure 3.3b; -0.850 m ≤ x ≤ -0.503 m WGS 84; -0.934 m ≤ x ≤ -0.614 m WGS

84; -0.930 m ≤ x ≤ -0.737 m WGS 84)

Figure 3.3: The interquartile range represents surface elevation envelopes per species of seedling/saplings surveyed in (a) aquaculture ponds and (b) reference forests Whiskers indicate maximum and minimum values and empty circles indicate outliers

3.3.2 Genera-specific surface elevation envelopes and prediction maps of mature mangrove trees in aquaculture ponds

The elevation envelopes at which trees occupied in reference forests are summarised in

Table 3.2 and Figure 3.4, and were derived from 59 Avicennia trees, 40 Rhizophora trees, 38 Sonneratia trees, and 2 Excoecaria trees Of all the tree species surveyed in reference forests, the surface elevation envelope occupied by Sonneratia spp was the

widest at 0.924 m ≤ x ≤ 0.598 m WGS 84, with second being Rhizophora spp of

Figure 3.4: The interquartile range represents surface elevation envelopes per genus (i.e

Avicennia spp., Excoecaria spp., Rhizophora spp and Sonneratia spp., surveyed in reference

forest sites Whiskers indicate maximum and minimum values and empty circles indicate outliers

Across the 29 aquaculture ponds, surface elevation was within the range of -2.158 m to 2.411 m WGS 84 (Figure 3.5a) Surveyed vegetation in aquaculture ponds have largely established at or below MSL and were more restricted to inter-tidal positions with higher elevations, within the range of -1.511 m ≤ x ≤ 0.228 m WGS 84 Of these 651 individuals, 137 individuals (21%) had established on regarded areas (Figure 3.5)

Figure 3.5: Map of aquaculture ponds showing surface elevation changes (i.e grade down, grade up), location of pile of broken branches and established vegetation where each green triangle represents an established individual (surveyed in June 2014)

The DEM in Figures 3.6a – 3.6d have been spatially classified into three categories – below zero (-2.152 m ≤ x < 0 m WGS 84), zero and above zero (0 m < x ≤ 2.411 m WGS 84) The green areas serve to represent potential, suitable elevation range

wherein Avicennia spp., Rhizophora spp And Sonneratia spp may establish in the future as trees and are reported to be 2.95% for Avicennia spp., 9.80% for Rhizophora spp., and 12.0% for Sonneratia spp of the total area

3.4 Discussion

3.4.1 Surface elevation affects propagule establishment and seedling development

The rehabilitated aquaculture ponds exhibited topographic heterogeneity, with elevations that ranged from -2.158 m ≤ x ≤ 2.411 m WGS 84 Post-rehabilitation surveys highlight that a total of 471 seedlings/saplings and 180 trees, across 13 species, had established in the aquaculture ponds (Table 3.1) The majority of seedlings/saplings surveyed in aquaculture ponds had established at more restricted inter-tidal positions with the range -1.511 m ≤ x ≤ 0.228 m WGS 84 (Figure 3.3a), compared to the full elevation range represented in ponds In general, established vegetation was restricted to the perimeter of aquaculture ponds (Figure 3.5b) where relatively higher surface elevations exist compared to the lower elevations found in the middle of ponds

Higher surface elevations around the pond perimeters relates to a lower inundation hydroperiod Inundation hydroperiod encompasses the frequency and duration a location is inundated, and is determined by surface elevation, tidal frequency and amplitude (Crase et al., 2013) Hence, inundation hydroperiod and its inherent link to surface elevation change is a key control on mangrove establishment, forest structure and subsequent long-term stability (Kitaya et al., 2002; Friess et al., 2012) Inundation hydroperiod was first proposed as a key control on mangrove communities more than

80 years ago (Watson, 1928) The distribution of mangrove communities could be delineated based on their tidal regime, surface elevation and inundation frequency This influence of hydroperiod on mangrove distribution has been examined more

recently in other field studies Crase et al., (2013) found that S alba trees dominated