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 6

Chapter – General Discussion Mangrove rehabilitation efforts are strongly supported by many governments, intergovernmental agencies and NGOs, but have met with mixed success. Since ecosystems are inherently dynamic, the ability to predict and manage rehabilitation trajectories are very challenging (Anand & Desrochers, 2004; Ruiz-Jaen & Aide, 2005). Palmer et al. (1997) and Hildebrand et al. (2005) cautioned that these complex landscapes should be partitioned into “understandable subsets” in order to develop theory and generalities that can further the understanding of “restoration ecology” (the equivalent of “rehabilitation ecology”). Furthermore, rehabilitation goals must be realistic and include multiple endpoints that can be subjected to scientific testing (Hildebrand et al., 2005). This thesis aimed to contribute to the understanding of hydrologic management in the success of mangrove rehabilitation projects. It aimed to investigate how the knowledge of (i) surface elevation in controlling mangrove establishment and development and (ii) species-specific inundation thresholds contributes to mangrove rehabilitation success. This was achieved through a field study and a controlled mesocosm experiment. 5.1 Principal findings 5.1.1. Effects of surface elevation on mangrove establishment The first objective investigated how hydrologic restoration contributed to the successful establishment of mangrove vegetation in a rehabilitated site. Overall, establishment rates were low and were restricted to higher elevation ranges around the perimeter of aquaculture ponds. Also, the surface elevations at which seedling/sapling 66 establishment occurred were similar across both rehabilitation site and natural reference forests. This highlights that surface elevation is a potentially a key influence on the establishment and colonisation of mangroves as it controls inundation hydroperiod (i.e. inundation frequency and duration). At inappropriate surface elevations, establishment of propagules becomes impeded through (i) lack of inundation-free periods and (ii) a sufficiently long inundation-free duration that favours both propagule stranding and root development for anchoring into substrate. Assuming successful establishment, surface elevations continue to influence subsequent development into seedlings and saplings. Effectively, the relationship between survival and growth of mangrove seedlings is that prolonged inundation hydroperiod (i.e. low elevations) decreases seedling survivorship and growth (Ellison & Farnsworth, 1997; Kitaya et al., 2002; Chen et al., 2005; Chen et al., 2013). Contributing reasons are that prolonged inundation impedes aerobic metabolism (McKee, 1996) and also causes a build-up of phytotoxins in anoxic soil that in turn affects plant growth through inhibition of photosynthetic gas exchange and oxygen available to roots (McKee, 1993; Youssef & Saenger, 1998). In mature mangrove communities, inundation hydroperiod has also been shown to explain for distribution patterns, first by Watson (1928) and more recently by Crase et al. (2013). 5.1.2. Effects of prolonged inundation on seedling survival and development The second objective quantified the inundation thresholds impacting the survival and growth for two selected mangrove species – Rhizophora mucronata and Avicennia alba. Analyses revealed species-specific responses to inundation duration. Avicennia seedlings exhibited an inundation threshold of approximately hours, after which 67 prolonged inundation decreased survivorship and inhibited seedling height. No such threshold was exhibited in Rhizophora seedlings. This species-specific response may be explained by the superior ability of Rhizophora to maintain positive net photosynthesis through sustaining functional photochemical and biochemical systems when (Pezeshki et al., 1987). This is accounted for by inter-species differences to in plant physiology (root aerenchyma to store oxygen), propagule reserves (Smith & Snedaker, 2000) and reproductive strategy (Tomlinson, 1986; Elmqvist & Cox, 1996; Tomlinson & Cox, 2000). 5.2 A synthesis: Reconciling a field study and a mesocosm experiment After reproductive propagules have escaped losses from pre-dispersal herbivory and dispersal, they have to overcome multiple constraints to establishment and early development of mangrove seedlings. These constraints either act in isolation or synergistically, encompassing variables such as temperature, carbon dioxide, salinity, light, nutrients, inundation and biotic entities (sensu Krauss et al., 2008). The field study presents field data that show establishment is constrained by surface elevations as mangrove establishment occurs within specific elevation ranges in both aquaculture ponds and natural reference forests. The field data was further expanded on by through executing a controlled mesocosm experiment, designed to provide complementary knowledge by determining species-specific thresholds to manipulated inundation durations. Additionally, the mesocosm experiment functioned to explicitly control for confounding factors in the field study which may exert a combined influence on seedlings. For instance, inundation tolerance thresholds were influenced by a salinity gradient in Kandelia obovata trees. K. obovata exhibited longer inundation duration tolerances as salinity decreased and this determined the spatial distribution of mature trees (Yang et al., 2013). The experiment investigated only the relationship between 68 manipulated inundation durations and its effect on seedling survival and growth. Thus observed responses can be directly linked to the influence of prolonged inundation, conditions similar to lower surface elevations in mangroves. Yet, the mesocosm experiment arguably over simplifies realistic conditions in the field. As mentioned above, the spatial distribution of mangrove establishment can be influenced by many factors. Hence, the need for a field validation to show that effect of surface elevation on establishment can still be observed even in face of a combination of such interacting factors. The field study further expands knowledge beyond two species as the data collected corresponds to the spatial distribution of mangroves across 14 species and life stages (seedling, sapling and trees). Taken together, this thesis provides an insight on the importance of appropriate surface elevations in establishment and colonisation for the reforestation of rehabilitated sites. Mangroves establish at specific surface elevation ranges and arguably, they can only survive and develop at optimal inundation durations upon successful establishment. It highlights the need for rehabilitation planners and practitioners to alter surface elevations in degraded sites to a degree that favours natural colonisation and establishment. 5.3 Implications for mangrove rehabilitation Managed appropriately, mangrove rehabilitation through natural succession or humanaided secondary succession can produce self-sustaining ecosystems resilient to normal periodic stresses (Simenstad et al., 2006; Elliott et al., 2007; Borja et al., 2010), It is encouraging to note that there exist rehabilitation projects that have employed this EMR approach in achieving successful rehabilitation (reforestation). Examples of 69 projects combining different hydrological and EMR approaches have been comprehensively summarised and discussed in an EMR manual published by Lewis and Brown (2014), Sections 8.7 and 10. Table 5.1 presents a comparison of the techniques employed in these projects in achieving successful mangrove reforestation. Each project employed a unique combination of EMR techniques, dependent on sitespecific requirements. Although two projects did involving replanting of mangroves, rehabilitation techniques employed in remaining projects proved sufficient in inducing natural colonisation of mangrove vegetation. In general, the techniques employed to rehabilitate aquaculture ponds in Kuri Caddi (i.e. strategic breaching, regrading and broadcasting of propagules) was similar to these projects. However, the rehabilitation goal of reforestation was not achieved due to inappropriate surface elevations in aquaculture ponds. 70 Table 5.1: A summary of the EMR rehabilitation techniques employed across eight mangrove rehabilitation projects in achieving successful reforestation (Lewis & Brown, 2014). Techniques U.S.A. 500 √ U.S.A. 1.7 √ 40 Dredged Material Islands, East and West Florida Coasts Tiwoho, North Sulawesi U.S.A. 108.5* √ (Fill was required) Indonesia 20 √ (Infilling of aquaculture channels) √ √ Tanakeke Island, South Sulawesi Indonesia 400 √ √ √ NE Langkat Wildlife Sanctuary, North Sumatera Indonesia 12 Hillsborough Bay, Florida U.S.A. Pelican Island, Florida U.S.A. Tidal creek Excavation Planting West Lake, Florida Sunken Island, Florida Regrading Area (Ha) Strategic Breaching Country Hydrological rehabilitation Name of site √ √ (Marsh grass planted to stabilise sediments) √ (Some mangrove replanting by the government and students √ Soil mounds to raise elevation √ √ (6 species were planted in half of entire area) √ √ (Marsh grass planted to stabilise sediments) √ (Marsh grass planted to stabilise sediments; Rhizophora mangle planted) *Additional data was retrieved from Lewis & Lewis, 1978. An immediate implication of this study is that mangrove rehabilitation practitioners should more strongly consider surface elevations as a critical component of mangrove 71 rehabilitation approaches. This thesis has provided evidence that surface elevation and its control on inundation hydroperiod has influences on the establishment, colonisation, survival and development of mangroves (Chapters and 4), and hence rehabilitation (reforestation) success. The thesis provides support that strengthens the main guiding principle of EMR whereby rehabilitation should work within the physical boundary conditions that control mangrove establishment and survival, such as tidal hydrology, flushing and wave action (Lewis, 2005; Lewis & Gilmore, 2007; Chen et al., 2012; Winterwerp et al., 2013, Lewis & Brown, 2014). Secondly, this study reiterates the importance for periodic monitoring and management of rehabilitation sites in achieving rehabilitation success. While there is initial success in seedling establishment and colonisation, continual monitoring is necessary to track future colonisation and forest regeneration trajectories for a complete assessment of the progress and success of the rehabilitation program. Rehabilitation and management should be constantly re-evaluated based on actual ecosystem response to management (Holling, 1978; Walters, 1986, Lewis & Brown, 2014). Lastly, there is an urgent need to get knowledge and data gleaned from such rehabilitation out to both rehabilitation scientists and practitioners. Only a small subset of rehabilitation projects implemented have been planned or analysed by ecologists and other scientists (Kentula, 2000). There is a general lack of data and consistent documentation describing rehabilitation efforts, making it difficult to review and determine the reasons for success or failure of rehabilitation projects. This impedes future efforts in improving rehabilitation design in achieving higher rates of rehabilitation success. 72 5.4 Recommendations 5.4.1. Identifying disturbance-free periods that favour establishment and colonisation Chapter provided evidence that inappropriate surface elevations were impeding successful mangrove establishment and colonisation. Yet, the nature of field studies encompasses the inability to control for other physical (e.g. similar substrate composition), biological (e.g. propagule fruiting seasons) and ecological factors (e.g. propagule predation by crabs; Smith et al., 2009) that could have influenced propagule establishment, colonisation and development processes. This was addressed partially through the mesocosm experiment that kept all other such variables constant and investigated solely, the relationship between seedling survival and development and inundation durations. An extension would be to investigate the relationship between inundation hydroperiod (i.e. inundation frequency and duration) and species-specific propagule establishment, across an elevation gradient. This was not viable in this study given logistical and funding restrictions. Yet, the knowledge gleaned would be useful in identifying the minimum inundation hydroperiod that would provide disturbancefree periods to allow sufficient time for root development in stranded propagules to withstand disturbances such as removal by tidal flooding and wave energy. Balke et al. (2011) coined the term “Windows of Opportunity” to describe these disturbance-free periods. It follows that such windows would vary across an elevation gradient, environmental conditions and between ecosystems. In the context of revegetation at degraded sites, a combination of such knowledge could guide rehabilitation practitioners in identifying the appropriate surface elevations and time in the tidal regime for such revegetation actions (e.g. broadcasting of mangrove propagules). 73 5.4.2 Implement mid-course corrections The vegetation survey for established vegetation in rehabilitated ponds was conducted in June 2014, months after rehabilitation was completed. Despite recognition in June 2014 that surface elevations in aquaculture ponds were too low, funding restrictions imposed by collaborative partners hindered the planned implementation of mid-course corrective actions. An ideal extension to improve on this study would be to test if increase in surface elevation and presence of more aboveground structures could aid propagule establishment. More dike walls could be graded down to produce substrate to increase the surface elevations in ponds to that similar in reference forests. Then, more aboveground structures such brush piles could be created to increase “surface roughness” before repeating the broadcast of propagules. Lastly, a second vegetation survey could be conducted to monitor if any new vegetation had established on site. 5.4.3 Long-term monitoring of recovery trajectory To predict if the rehabilitation goal of reforesting aquaculture ponds can be achieved in a few years, prediction maps were generated. Surface elevations of established trees in natural reference forests were used to identify comparable surface elevations in aquaculture ponds. Yet, it is recognised that the predicted distributions may be overly conservative as it was founded on a subset, instead of the full surface elevation gradients that natural mangrove forests can occupy. Effectively, the prediction maps created in this study fail to account for the upper elevation range (on the landward side) that mangroves once occupied before conversion into aquaculture ponds, agricultural fields and other urban developments. Also, while the Rhizophora seedlings had survived the experiment duration of two months, they might eventually die when energy stores within the propagule finish. Instead of attempts at predicting the 74 rehabilitation outcome based on reference forests and through a mesocosm experiment, the focus should be shifted into investing effort on conducting periodic monitoring of reforestation processes. This would allow – (i) early identification of unfavourable environmental conditions towards natural establishment and colonisation, and (ii) immediate implementation of mid-course corrective actions to mitigate these issues. 5.6 Conclusions Mangrove rehabilitation efforts globally have experienced mixed success, as knowledge of species-specific inundation thresholds have not been incorporated into rehabilitation design. This is despite the fact that the role of surface elevation and inundation frequency in controlling mangrove establishment and stability has been known for decades (i.e. Watson, 1928; Crase et al., 2013). Moreover, mangrove research is commonly executed in isolation from the needs of rehabilitation practioners, and there is an urgent requirement for applied ecological research aimed at testing and critically analysing the decisions made in rehabilitation projects. This will produce data that can be used towards refining rehabilitation designs in achieving rehabilitation success. Through relating surface elevation thresholds and its control on inundation hydroperiod, to vegetation recruitment dynamics, this study provides empirical information of such thresholds that will assist practitioners in optimising the establishment and survival rates of recruited vegetation. Science-informed restoration that acknowledges geomorphological-ecological thresholds will contribute to the better restoration of this imperilled coastal ecosystem. 75 . that surface elevation and its control on inundation hydroperiod has influences on the establishment, colonisation, survival and development of mangroves (Chapters 3 and 4), and hence rehabilitation. surface elevation is a potentially a key influence on the establishment and colonisation of mangroves as it controls inundation hydroperiod (i.e. inundation frequency and duration). At inappropriate. linked to the influence of prolonged inundation, conditions similar to lower surface elevations in mangroves. Yet, the mesocosm experiment arguably over simplifies realistic conditions in the