Environmental Monitoring and Management of Reclamations ... · ENVIRONMENTAL MONITORING AND MANAGEMENT OF RECLAMATIONS WORKS CLOSE TO SENSITIVE HABITATS STÉPHANIE M. DOORN-GROEN

ENVIRONMENTAL MONITORING ANDMANAGEMENT OF RECLAMATIONS WORKSCLOSE TO SENSITIVE HABITATS

STÉPHANIE M. DOORN-GROEN

ABSTRACT

Traditional methods for environmentalmanagement of marine reclamation worksclose to sensitive habitats have generallynot provided the level of control necessaryto ensure preservation of these habitats.Obtaining the level of control necessary toassure authorities and non-governmentalorganisations (NGOs) of compliance withenvironmental quality objectives, requiresquantifiable compliance targets coveringmultiple temporal and spatial scales.

Of equal importance to allow feedback ofmonitoring results into compliance targetsand work methods are effective and rapidresponse mechanisms. This article describesthe successful implementation ofcomprehensive Environmental Monitoringand Management Plans (EMMP), basedupon such feedback principles, which allowreclamation activities to proceed in closeproximity to Singapore’s most importantmarine habitats under third party scrutiny.

Specific focus is placed on describing themethods utilised to quantify compliancewith daily spill budget targets and howsuch targets and compliances are assessed.To improve reliability, the spill budgets take

into account specific habitat tolerance limitsfor varying magnitudes and durations ofsediment loading. Refinements to sedimentplume models were undertaken to enhancetheir ability to hindcast impacts from thecontractors’ complex reclamation schedules.Methods for segregation of impacts andassessment of cumulative impacts were alsointegrated into the hindcast procedures.Finally, the article describes the updating of tolerance limits and confirmation of spillbudgets via targeted habitat monitoring.

To date, the EMMPs have been able todocument compliance of the works to allpre-project environmental quality objectivesat a level of reliability that cannot be refutedby third parties. This has minimised thedevelopers’ and contractors’ exposure topublic complaints and liabilities associatedwith environmental impacts. The EMMPshave thus allowed the reclamation activitiesto proceed in an efficient manner, whilstensuring protection of the environment.

The author wishes to acknowledge theimportant contributions of Thomas M. Foster,

Regional Director Southeast Asia, DHI Water& Environment (S) Pte Ltd to this research.This paper was first presented at WODCONXVIII in June 2007 and was published in theconference Proceedings. It is reprinted herein a slightly revised version with permission.

INTRODUCTION

The tropical waters in Singapore provideexcellent conditions for marine life, owingto relatively constant tropical watertemperatures and frequent fresh oceanthrough flow from both the South ChinaSea and Melaka Straits. Coral, seagrass andmangrove habitats have been found to berelatively rich in Singapore. The diversity ofthe coral habitats in Singapore is confirmedby the fact that of the 106 coral generaexisting world wide (Veron et al. 2000), 55 genera are documented in Singaporewaters alone (Tun et al. 2004), compared to 13 genera found in the Caribbean. For seagrass habitats, 12 species out of 57known species are found in Singapore(Waycott et al. 2004), whereas 24 out of54 true and minor mangrove species havebeen found in Singapore so far (Thomlinson1999). These numbers document the highdiversity of marine habitats in a relatively

Above, For coral reef areas subject to direct impact,

coral relocation is undertaken prior to the start of

reclamations works.

Environmental Monitoring and Management of Reclamations Works Close to Sensitive Habitats 3

4 Terra et Aqua | Number 108 | September 2007

small environment as Singapore andemphasize the importance of marinehabitat conservation in Singapore.

Owing to the confined nature of Singaporewaters and the presence of a large numberof patch reefs, reclamation and associateddredging activities (in the following referredto generically as reclamation activities),often take place in very close proximity tocoral reefs and seagrass areas. In addition,increasing industrial development results in developments also occurring close tosensitive industrial water intakes.

Recognizing the value of these marinehabitat and industrial resources, Singaporehas established strict Environmental QualityObjectives (EQO) for marine constructionactivities. In order to document compliancewith these EQOs, pro-active EnvironmentalMonitoring and Management Plans (EMMP)based upon feedback monitoring principlesare required for marine construction activitiesto proceed, when these are in closeproximity to key environmental receptors.

Introduced in Europe in the 1990s andrefined during the EMMP works for theØresund Link between Denmark andSweden (Møller 2000) and Bali Turtle Island,Indonesia (Driscoll et al. 1997), feedbackEMMP provides the level of responsivenessand documentation necessary to assureboth authorities and other interest groupsthat the works meet the EQOs throughoutthe construction period.

Based upon the strict nature of the EQOs,EMMPs in Singapore are required toestablish compliance of the works acrossmultiple temporal and spatial scales:• Compliance assessment against daily

spill budget targets at the work area;• Real-time monitoring and compliance

assessment against response limits,particularly for intakes and reefs in close proximity to the work area;

• Compliance assessment against resultsof daily hindcast modelling compared to habitat tolerance limits throughoutthe potential impact area.

The feedback mechanism allows forupdating of the spill budget limits, response

limits and tolerance limits, based on theresults of sedimentation monitoring andhabitat monitoring. To ensure the accuracyof the entire system, the performance isconfirmed on a daily basis via controlmonitoring of sediment spill.

This article presents how the variouscomponents of the EMMP are established andexecuted, together with the refinementsnecessary to ensure a level of responsivenessappropriate to the importance of thereceptors.

BASIC COMPONENTS OF THE EMMP

The EMMP is the primary method of controlto ensure EQOs relating to marine habitatsand other environmental receptors are met.The EMMP is further a tool to:• detect any unexpected impacts at an

early stage, • establish the response necessary to

address such impacts, and • confirm that appropriate tolerance

limits have been adopted.

The feedback approach of the EMMP ispro-active. It links the results of detailednumerical hindcast models of the sedimentplumes resulting from reclamation activitieswith the results from online turbidity andcurrent sensors, daily spill measurementsand periodic habitat surveys, and comparesthese against the spill budget.

The spill budget is the maximum allowablespill (daily, weekly and fortnightly limits are set), which will still ensure (based onthe results of sediment plume forecastmodelling) that the EQOs are met. As environmental receptors, like corals,have a different tolerance againstsuspended sediment levels than for examplemangroves, individual tolerance limits aredefined for each environmental receptor.

The tolerance limits play an important rolethroughout the project, as the daily spillbudget is based on these individual limits.Both tolerance limits and spill budgets areevaluated and updated during the project,based on results of the habitat monitoringcampaigns.

The main components of Feedback EMMPs,as implemented in Singapore are:

i) Environmental BaselineFeedback variables are identified,instrumented and monitored for astatistically significant period prior to construction, which is typically in the order of 3 to 6 months. Variables monitored include all key environmental receptors such as corals reefs, seagrass beds,mangroves, turbidity, water quality,currents and sedimentation. This phase also includes theconfirmation of the environmentalquality objectives for the project and environmental tolerance limits. If compensatory works are required, such as coral relocation from directimpact areas, this is also undertaken at this stage of the EMMP.

ii) Elaboration of work plansThe appointed reclamation contractorelaborates a work plan, specifying thedistribution of the work in time andspace, procedures and equipment.

iii) Assessment of work plansThe effect of performing the work planon the environment is assessed throughthe use of numerical sediment plumeforecast modelling.

vi) Revision of work plans If the forecasted impact resulting fromimplementation of the work plan leadsto unacceptable effects, i.e. violation of EQOs, the work plan is revised andreassessed. Once the work plan isfinalised a final EMMP specificationdocument is drawn up that specifies the detailed execution, response andmanagement process for the EMMP. In particular, the final EMMP specificationincludes a spill budget (for each phaseof the reclamation), which is the limitingamount of spill that will still result in theEQOs being met and against which theday-to-day control of the reclamationwork can be assessed.

v) Construction phaseReclamation commences.

vi) Compliance monitoringMonitoring of daily compliance variablesagainst the pre-determined sedimentspill limits (spill budget). If dailycompliance limits are violated, mitigationactions are established and implemented.If no violations of limits occur, reclamationwork and daily monitoring continue.Compliance monitoring is reported on a daily basis and to ensure the level ofresponsiveness, reporting is required amaximum of 45 hours in arrears of anyreclamation activity.

vii) Control monitoringMonitoring of real time measurementsand comparison to response limits, such as online turbidity data or weeklysedimentation data. If no violations oflimits occur, work and control monitoringcontinue. Control monitoring is reportedto the time scale of the monitoringactivity (daily or weekly).

viii) Spill hindcastSpill hindcast documents the impact of the reclamation progress on theenvironment remote to the work site.The spill hindcast is based upon realizedproduction schedules, composition offill material and actual tide conditions.The assessment is made through theuse of numerical sediment plumehindcast modelling, with the hindcastupdated every day. Reporting of thehindcast is made a maximum of threedays in arrears of the actual progress ofthe reclamation works so that remoteimpacts are captured prior to thembecoming significant.

ix) Habitat monitoringMonitoring of biological habitatfeedback variables is performed to anappropriate time schedule for theanticipated response rates. This is typicallyonce every three months for coral reefs,seagrass beds and mangrove areas.

x) Evaluation of construction phaseBased on the results of the biologicalmonitoring of feedback variables andthe results of the numerical spillhindcast modelling of the realizedconstruction process, the temporal and

IADC AWARD 2007

PRESENTED AT WODCON XVIII,

ORLANDO, FLORIDA, USA

MAY 27-JUNE 1, 2007

An IADC Best Paper Award was presented to

Stéphanie M. Doorn-Groen, Manager Engineering

Services at DHI Singapore, who has been based

in Southeast Asia since May 2002 and joined DHI

Singapore in January 2004. She graduated in

2000 with a BSc (Civil Engineering) from the

Polytechnic The Hague, the Netherlands and in

2002 with a MSc (Civil Engineering Management

& Geotechnology) from South Bank University

London, UK. Her previous experience was as a

geotechnical adviser for Fugro Onshore

Geotechnical bv, as a superintendent and

technical employee for the dredging company

Van Oord bv and for Municipal Works, Ports,

Design & Construct, Rotterdam, the Netherlands.

Each year at selected conferences, the

International Association of Dredging Companies

grants awards for the best papers written by

younger authors. In each case the Conference

Paper Committee is asked to recommend a

prizewinner whose paper makes a significant

contribution to the literature on dredging and

related fields. The purpose of the IADC Award is

“to stimulate the promotion of new ideas and

encourage younger men and women in the

dredging industry”. The winner of an IADC Award

receives Euros 1000 and a certificate of

recognition and the paper may then be published

in Terra et Aqua.


spatial impacts of the constructionphase are assessed. If EQOs are violated,mitigation actions are established,assessed and undertaken. On the basisof the realized impacts, environmentalcriteria (tolerance limits) and compliancecriteria (spill budgets) for the nextconstruction phase are updated (thefeedback loop).

xi) Next construction phaseThe construction and monitoring processreturns to task v) for each major stageof the reclamation and the process isrepeated until reclamation is complete.

xii) Completion of constructionAn environmental audit is produced atthe end of the construction period asformal documentation of the impactsrealised during the construction phase.This is based upon the result of thecompliance, control, habitat and supportmonitoring together with the results ofthe hindcast modelling of impacts. Theenvironmental audit is based on a finalhabitat survey usually carried out threemonths after the end of construction.

The main advantages of this approach toEMMPs are:• Compliance measurements are targeted

in the sediment plume resulting fromdredging and reclamation activities, as close as possible to the source of spill at the given time of measurement.This provides a much more accuratemeasurement of suspended sedimentspill than can be achieved via fixedturbidity sensor stations, which often lie outside the sediment plume forindividual dredging or reclamationoperations.

• Numerical sediment plume forecastmodels allow assessment of changes to the spill budget for variations incomplex reclamation schedules andvarying tide and ocean currentconditions, thereby ensuring the spillbudget is the most appropriate for thegiven stage of the works given thespecific equipment to be utilized andtiming of the activity.

• The hindcast model documents thespatial distribution of impacts at all

Stéphanie M. Doorn-Groen receiving an IADC

Award for young authors from Constantijn

Dolmans, Secretary General of IADC.


the receptor sites in the vicinity of thereclamation site with far broader spatialscale and finer temporal resolution thancan be achieved via habitat monitoringin isolation.

• The hindcast model keeps a runningbalance of the cumulative sedimentationimpact levels based on actual productionprovided by the reclamation contractor.Increasing levels of sedimentation can bedetected at an early stage and mitigatingmeasures can be applied, if necessary.

• The combined use of daily spillcompliance monitoring, controlmonitoring and hindcast modelling allowsthe EMMP to respond rapidly and reliablyto different temporal impact scales (fromfor example, short term exceedencesresulting from, for example, unexpectedevents, to long-term trends resultingfrom, for example, deterioration in thequality of fill material).

• The feedback loop ensures that tolerancelimits and resultant spill budgets areconsistent with the specific sensitivity of the environmental receptors in theimpact area.

ENVIRONMENTAL QUALITY OBJECTIVES

In order to set EQOs for a project, it isessential that a classification scale is adoptedto define the scale of impacts that may beallowed at a given environmental receptor.The following scale of impact classificationshas been adopted for several projects inSingapore:• No impact: Changes are significantly

below physical detection level andbelow the reliability of numericalmodels, so that no change to thequality or functionality of the receptorwill occur.

• Slight impact: Changes can be resolvedby numerical sediment plume models,but are difficult to detect in the field asthey are associated with changes thatcause stress, not mortality, to marineecosystems. Slight impacts may berecoverable once the stress factor hasbeen removed.

• Minor impact: Changes can be resolvedby the numerical models and are likelyto be detected in the field as localized

mortalities, but to a spatial scale that is unlikely to have any secondaryconsequences.

• Moderate impact: Changes can beresolved by the numerical models and aredetectable in the field. Moderate impactsare expected to be locally significant.

• Major impact: Changes are detectablein the field and are likely to be relatedto complete habitat loss. Major impactsare likely to have secondary influenceson other ecosystems.

The task of defining EQOs rests with theauthorities and is made on an area by area,habitat by habitat basis. For reclamationprojects in Singapore, “Slight Impact” istypically allowed in the area immediatelyadjacent to (within 500 m) of the workarea, whilst “No Impact” is required for allenvironmental receptors remote from thework area. For coral reef areas subject todirect impact (i.e. under the reclamationprofile), it is presently common practice inSingapore to compensate for the habitatloss by undertaking a coral relocationexercise prior to start of reclamation works.

TOLERANCE LIMITS ANDENVIRONMENTAL QUALITY OBJECTIVES

The linkage between project EQOs and spill budget depends on the method ofreclamation and on the tolerance limits ofthe various environmental receptors, whichin turn depends upon the pre-projectexternal stress levels on the ecosystem. In Singapore, initial tolerance limits for the most sensitive marine habitats (coralsand seagrass) have been established basedupon extensive literature review and DHI’sexperience from similar projects in theSouth East Asia region.

These tolerance limits have then beenrefined over the course of several projectsin Singapore, based upon the results ofproject specific habitat monitoring.Presently, these limits, as presented below,are believed to be the most relevant set oftolerance data available for coral reefs andseagrass beds subject to incrementalreclamation impacts on top of elevatedexternal (non-project related) stress levels.

Coral tolerance to suspendedsedimentsIn simplified terms, as hard corals aredependent on symbiotic photosynthesizingzooxanthellae for their nutrient supply andsurvival, they are sensitive to increasedturbidity levels as the reduction in lightpenetration through the water columnadversely affects the photosynthesisprocess. Perhaps more seriously, elevatedsedimentation levels can clog the corals’respiratory and feeding system, whilst alsocausing complete light extinction to theimpacted area of the colony.

The level of sensitivity depends on thecharacteristics of the corals, with platecorals like Pachyseris sp. proving the mostsensitive to increased sedimentation andleast sensitive to reduction in lightpenetration. Conversely, branching coralssuch as Acropora sp. show the oppositesensitivity trend. Clearly, other impacts suchas degradation of substrate impactingattachment of coralline larvae are alsoimportant to the overall impact levelexperienced by a reef. However, the presentstate of the art cannot quantify suchdetails, which are therefore captured viathe habitat monitoring component of theEMMP rather than via the tolerance limits.Background levels vary from region toregion and are very site specific. Researchfrom the Barrier Reef (Harriot et al. 1988)indicates that these corals are tolerant tolevels of suspended sediments up to 4 mg/l(absolute concentration). However, studiesin Hong Kong have shown tolerance levelsup to 10 mg/l. Extensive monitoring datafrom multiple projects in Singapore (wherechanges in reef health, measured as afunction of live hard coral cover anddiversity, has been compared to measuredand predicted suspended sediment andsedimentation levels), has allowed thedevelopment of an coral tolerance matrixfor excess (above background) suspendedsediment concentrations, see Table I.

This table is found to be applicable for theelevated background turbidity levels commonin Singapore and the typical Singapore reefmorphology, which is dominated by themore resilient massive corals and platecorals, as shown in Figure 1.

Coral tolerance to sedimentationCoral sensitivities to sedimentation aredetermined largely by the particle-trappingproperties of the colony and the ability ofindividual polyps to reject settled materials(Figure 2). Horizontal plate-like colonies andmassive growth forms present large stablesurfaces for the interception and retentionof settling solids. Conversely, vertical platesand upright branching forms are less likelyto retain sediments.

A threshold (absolute) value of 0.1 kg/m2/dayhas previously been adopted as the criticalvalue for corals in Environmental ImpactAssessments in Hong Kong. However,monitoring data from Singapore indicatesthat an incremental value of 0.05 kg/m2/dayis more appropriate for the type of coralhabitats and existing stress levels in Singaporewaters. Based on these Singapore data sets,the tolerance limits presented in Table II arefound to be relevant for sedimentationimpact on corals for reefs with naturally highbackground sedimentation levels, assuming a net deposition density of 400 kg/m3.

Seagrass tolerance to suspendedsedimentsProductivity of seagrass can be limited owingto reduced light penetration resulting fromthe presence of algal blooms and suspendedsediments. Seagrass requirements for lightpenetration have been well described bymultiple authors, with the habitat beingconfined to water depths where light levelsare above 10% to 15% of surfaceirradiance. For the normal tidal rangeexperienced in the Singapore area, these

figures concur well with observations withinSingapore waters, which indicates thatseagrass are generally limited to seabedareas shallower than –1 m CD. At low tide,many seagrass beds in the Singapore area

are exposed (Figure 3), indicating a possibleadaptation to, and higher tolerance to ahigh level of excess suspended sediments.For deeper beds, the tolerance is lower, butgiven the natural background variability in

Figure 1. Typical coral habitats in Singapore.

Table I. Impact severity matrix for suspended sediments on corals inenvironments with high background concentrations

Severity Definition (excess concentration)No Impact Excess Suspended Sediment Concentration > 5 mg/l

for less than 5% of the time

Slight Impact Excess Suspended Sediment Concentration > 5 mg/l


Excess Suspended Sediment Concentration > 10 mg/l


Minor Impact Excess Suspended Sediment Concentration > 5 mg/l

for more than 20% of the time



Moderate Impact Excess Suspended Sediment Concentration > 10 mg/l




Major Impact Excess Suspended Sediment Concentration > 25 mg/l




Table II. Impact severity matrix for sedimentation impact on corals

Severity Definition (excess sedimentation)

No Impact Sedimentation < 0.05 kg/m2/day (<1.7 mm/14 day)

Slight Impact Sedimentation < 0.1 kg/m2/day (<3.5 mm/14 day)

Minor Impact Sedimentation < 0.2 kg/m2/day (<7.0 mm/14 day)

Moderate Impact Sedimentation < 0.5 kg/m2/day (<17.5 mm/14 day)

Major Impact Sedimentation > 0.5 kg/m2/day (>17.5 mm/14 day)

suspended sediment load in the Singaporearea, it is reasonable to assume that theouter limits of the seagrass are welladapted (in terms of water depth) to short-term fluctuations in the backgroundconcentration of 5 to 10 mg/l, such thatexcess loadings higher than 5 mg/l will berequired to stimulate a noticeable habitatchange. These findings, coupled withmonitoring experience from the SE Asiaregion, result in the proposed impactseverity matrix presented in Table III.

Seagrass tolerance to sedimentationThe growth rates of seagrass are high.Growth in the order of 1 to 2 cm per dayhas been recorded for example forThalassia sp. (Durate et al. 1999) whilstgrowth rates in the order of 0.6 cm per dayhave been recorded for Enhalus sp. in Malaysia.Therefore, the short-term survival ofseagrass beds, which depends on anaerobicperformance, will only be impacted in thecase of very high sedimentation rates. Suchcritical sedimentation rates will normallyonly occur very close to a reclamation site.Based on experience in the SE Asia region,the following impact severity matrix ispresented for sedimentation impact onseagrass (see Table IV). Other impactsresulting from increased sedimentation,such as change in composition of substrate,are clearly also important to the overallimpact levels experienced by a seagrass bed,but such detailed impacts are difficult toquantify and are therefore captured via thehabitat monitoring component of the EMMP.

Figure 2. Sedimentation impact on corals and expulsion of sediment via mucus generation.

Table III. Impact severity matrix for suspended sediment impact on Seagrass in high background environments

Severity Definition (excess concentrations)No Impact Excess Suspended Sediment Concentration > 5 mg/l

for less than 20% of the timeSlight Impact Excess Suspended Sediment Concentration > 5 mg/l

for more than 20% of the timeExcess Suspended Sediment Concentration > 10 mg/l for less than 20% of the time

Minor Impact Excess Suspended Sediment Concentration > 25 mg/l for less than 5% of the time

Moderate Impact Excess Suspended Sediment Concentration > 25 mg/l for more than 20% of the timeExcess Suspended Sediment Concentration > 75 mg/l for less than 1% of the time

Major Impact Excess Suspended Sediment Concentration > 75mg/l for more than 20% of the time

Figure 3. Typical seagrass habitat

in Singapore.

Table IV. Impact severity matrix for sedimentation impact on Seagrass in highbackground environments

Severity Definition (Excess sedimentation)No Impact Sedimentation < 0.1 kg/m2/day (<0.25 mm/day)Slight Impact Sedimentation < 0.25 kg/m2/day (<0.63 mm/day)Minor Impact Sedimentation < 0.5 kg/m2/day (<1.25 mm/day)Moderate Impact Sedimentation < 1.0 kg/m2/day (<2.5 mm/day)Major Impact Sedimentation > 1.0 kg/m2/day (>2.5 mm/day)


Mangrove tolerance to suspendedsediments and sedimentationMangroves can be considered to be verytolerant to the range of suspendedsediment loads that may be generated from dredging and reclamation activities.Of the various mangrove species, thosewith pneumatophore root systems are themost sensitive to sedimentation (Thampanyaet al. 2002), but even mangroves withpneumatophore root systems are only likely to be stressed when prolongedsedimentation reach levels from 10 cm upto 30 cm. This level of sedimentation isunlikely to occur outside the work area,and mangroves are thus not considered assensitive receptors. Never-the-less, as EQOsare normally specified for mangrove areas,they are normally included in the habitatmonitoring campaigns for reclamationEMMP in Singapore. Figure 4 presents atypical mangrove habitat in Singapore area.

Visual impact and detection limitsIn the turbid environments that are foundaround Singapore, low concentrationsediment plumes in the surface of thewater column are generally not visible(based upon results of remote sensinganalysis) if the excess concentration abovebackground does not exceed 5 mg/l.A realistic measurable visual detection limitfor non-recreational areas (in the Singaporehigh background turbidity context) wouldbe a reoccurring plume present for 30-40 minutes per 12 hour daylight period,i.e. an exceedence of about 5% per day,whilst for recreation areas a limit of 2.5%exceedence proves to be appropriate.

Intake tolerance limits to suspendedsedimentsFor many industrial intakes the absolutetolerance limit to suspended sediments isnot known by the operators. In such cases,the most practical method for establishinga tolerance limit is to carry out statisticalanalysis on long-term background suspendedsediment data from the immediate area ofthe intake. It is then possible to carry out atest for no statistical change (at a confidencelimit agreed with the operator) for thevarious time scales of interest (daily, weekly,monthly and 6 monthly tests are normallyconsidered in Singapore).

DAILY COMPLIANCE MONITORING

Based on the EQOs, spill budgets aredefined for each stage of the reclamationworks. The spill budgets are updated aswork progresses and feedback informationconfirms their applicability or indicates arelaxation or tightening is warranted.

The contractor’s compliance to the daily spillbudget is assessed on a daily basis against dailyspill budget targets and on a weekly basisagainst weekly and fortnightly spill budgettargets. Typically, the fortnightly spill budget is60% of the daily spill budget, reflecting theability of most receptors to cope with higherlevels of stress if they are short-term or inter-mittent in nature. Daily compliance to spillbudget targets must be established within atime frame which will allow response beforeany non-compliance will pose a threat to theenvironment. Therefore, daily compliancemonitoring requires strict daily procedures fordata delivery from the contractor, to ensuredaily spill calculations, laboratory analysis dailycompliance analysis and reporting can becarried out in a timely manner.

On daily basis the contractor supplies:• Realised dredging volumes per dredger

for every single trip, including locationand method;

• Start and end time of dredging cycle,including delays;

• Realised reclamation volumes perdredger for every single trip, includinglocation and method;

• Start and end time of reclamation cycle,including delays;

• Representative sediment samples fromeach load to be analyzed for finecontents by an external laboratory.

In addition to the data provided by thecontractor, suspended sediment samplesare taken in the main plume discharge ofthe reclamation site (either as suspendedsediment samples (TSS) or sediment fluxmeasurements using acoustic backscattertechnology). The location and the time ofthe sampling reflect the reclamationactivities of the contractor and the samplesare analysed for TSS by an external laboratory.

This analysis provides a second method ofcontrol and serves as a validation for thespill calculation and performance of thenumerical hindcast models. Based upon thefines content of the fill and dredge materialand method of reclamation an empiricalestimate of the total spill is made, for eachreclamation/dredging operation over thepreceding 24 hr period. The resultant totalcan then be compared to the spill budget

Figure 4. Typical mangrove habitat in Singapore: Avicennia pneumatophore system (foreground) and

Rhizophora stilt root system (background).


on the level of compliance established forthe 24 hr period. A typical example is shownbelow for rainbowing operations.

Spill of fines leaving immediate dredgingarea = Load volume * fines % * 25%

Although it is recognised that the specificspill is dependent on many factors such asthe prevailing water depth and currentspeed, this simple empirical formula hasproved to be a reliable method forestimation of spill for sand placement inthe typical range of physical conditionsencountered in Singapore.

Validation of the spill calculation issubsequently provided by the TSS orsediment flux measurements in the plume.However, for the purpose of the dailycompliance monitoring a simple, yetreliable, empirical formulation is required to meet the reporting time scale.

Figure 5 presents a general flowchart of the complex daily operating proceduresrequired to establish compliance with spillbudget targets to a time frame which willallow response before any non-compliancewill pose a threat to the environment,which has been defined as a maximum of45 hrs in arrears of any activity on site.

Figure 6 presents an example of the dailyspill calculations over a period of fivemonths based on the empirical methodsdescribed above and validated by thecontrol sampling in the sediment plume.This figure indicates that the daily spillbudget was exceeded for a period in April. Mitigating measures were introducedand subsequently the spill budget wasachieved for the remainder of thereclamation work.

Figure 5. Daily operating procedure for

daily compliance monitoring.

Figure 6. Spill results from

reclamation operations.


DAILY HINDCAST MODELLING

Based upon the information provided bythe contractor in terms of time and locationof activities and the calculated spill, dailyspill hindcast simulations are run in order toestablish the temporal and spatial impactsof the sediment plumes released from thework area.

Hydrodynamic model setup andperformanceThe daily hindcast modelling is based uponDHI’s extensively verified 675/225/75/25 mMIKE 21 nested grid hydrodynamic modelof the Singapore Straits, which wasdeveloped in 2001 and is being continuouslyrefined on the basis of daily real time currentmeasurements.

Figure 7 shows the overall regional modelgrid coverage utilised for EMMP projects inthe Singapore area, whilst Figure 8 presentsan example of the model performancewhich meets relevant international standardssuch as UK Foundation for Water ResearchPublication Ref FR0374 “A framework formarine and estuarine model specification inthe UK”. The 25 m Model resolution isadopted in the specific area of reclamationto ensure all relevant local hydrodynamicfactors, which may affect the plumetransport and dispersion are resolved.

Bathymetric survey data are taken directlyfrom digital navigation charts,supplemented by project specific surveydata, which is updated on a weekly basisfor reclamation progress in the specificproject areas.

Figure 7. DHI’s Singapore Straits regional

675 m grid model with nested

intermediate 225 m grid and local

75 m grid sub-domain models.

Figure 8. Example performance of DHI’s Singapore

Straits current forecast model. RMS error on current

speed at presented validation point = 0.09 m/s.


Sediment plume model set-up andperformanceCalibration and validation of DHI’s sedimentplume hindcast model for Singapore watershas been carried out over the course ofseveral projects. A typical example of themodel performance is provided in Figure 9.Throughout the course of the EMMP, theperformance of the model is verified on adaily basis, either by direct TSS measurementswithin the sediment plume or via sedimentflux transects through the plume.

Critical shear stress for erosion anddepositionA vital factor to the performance of the modelin terms of documenting impacts on coral reefhabitats is the parameterisation of the criticalshear stress for erosion and deposition overthe reef areas. The complex morphology ofcoral reefs on both micro and macro scales,leads to an increased tendency for depositionto occur and a reduced tendency for re-suspension. Extensive testing andcomparison to sediment trap data collectedon a weekly basis has been undertaken,leading to the following conclusionsconcerning average critical shear stressparameters for deposition and re-suspensionof fines over coral reef areas in Singapore:

• Critical shear stress for deposition offine material over coral reef: 0.6 N/m2

• Critical shear stress for re-suspension ofinitial deposits over coral reef: 1.5 N/m2

Figure 10 presents the example maps ofcritical shear stress in South-WestSingapore, whilst Table V presents anexample of model performance againstmeasured sediment trap data.

Location survey vessel

Water samples were taken

TSS results:

Point 15: 2.5 mg/l

Point 16: 12.0 mg/l

Point 17: 40.0 mg/l

Point 18: 6.0 mg/l

Point 19: 1.8 mg/l

Figure 9. Location and magnitude of the sediment plume predicted by DHI’s hindcast model

and the location of the survey vessel during plume transects.

Table V. Example of model performance measured against reef sedimentation data

Figure 10. Maps of critical shear stress for erosion (left) and deposition (right) covering

SW Singapore for sediment released from dredging and reclamation operations.

Measured incrementalsedimentationKg/m2/day

0.020.04

Predicted incrementalsedimentation withoutadjustment of critical shear stress parameters

< 0.01< 0.01

Predicted incrementalsedimentation withadjustment of critical shear stress parameters

0.040.04


Sediment settling velocity In order to reliably simulate the transportand fate of the fine material released fromdredging and reclamation activities, it alsoproves necessary to divide the sedimentspill into a number of sediment fractions.After testing of various options, 6 fractions(3 for reclamation fill and 3 for dredgematerial) have been found to provide agenerally consistent compromise betweenmodel reliability and computational time,which is critical to the reporting schedule.

In order to establish the characteristics ofthe 6 sediment fractions, fall velocitytesting of the fine material present in thereclamation and dredge material is carriedout on a regular basis via Owen tube tests.Fall velocity characteristics are typicallyupdated on a monthly basis (separately forreclamation fill and dredged material), orwhen the daily control measurements inthe sediment plume indicate a necessity forupdating. An example of the Owen tubetest results is provided in Figure 11.

Execution of daily hindcastBased on the contractor’s activity informationand calculated spill, the numerical spillhindcast is carried out on a daily basis forthe actual reclamation operations. The result

of the daily EMMP hindcast model arevalidated against the daily control samplestaken in the sediment plumes originatingfrom the reclamation.

The daily hindcast is processed to allowdirect comparison to the EQOs with thefollowing key outputs:• Time series and tabulation of excess

suspended sediment concentration atthe various environmental receptors;

• Maps of exceedences of 5, 10 and 25 mg/l excess concentration;

• Animations of concentration maps.

UPDATING OF TOLERANCE LIMITS ANDSPILL BUDGET

As the spill budget is dependent on thetolerance limits of the various environmentalreceptors, it is critical that the reliability ofthese limits is confirmed at an early stageof the construction works, with continuousrefinement carried out throughout theconstruction period. The tolerance limits areconfirmed (or refined) based upon theresults of quarterly habitat monitoring ofkey environmental indicators compared tothe results of the sediment plume hindcastand sedimentation monitoring.

Habitat monitoringQuarterly control habitat monitoring surveysare carried out to establish the status of thevarious marine habitats near the developmentsite. The choice of survey locations is basedupon three criteria:• Importance and/or sensitivity of the

habitat;• Expected level of impact (based upon

the sediment plume forecast); and• Control stations outside the potential

impact area (based upon the sedimentplume forecast).

For each survey station key indicators areidentified and the survey sites laid out tofacilitate exact replicate surveys.

Coral habitat monitoringCoral surveys are primarily carried out usingthe Line Intercept Transect (LIT) method, asshown in Figure 12, which is recommendedby the Global Coral Reef Monitoring Network(English et al. 1997, Hill et al. 2004) forquantification of the percentage cover ofreef building corals, coral diversity, as wellas other benthic life forms. The LITmethodology, which provides a goodmethod for identification of mortalities oflarger reef areas, is supplemented by exactrepeat surveys of selected individual colonies,

Fraction 1 60% contribution

Representative fall velocity v = 0.00075 m/s

Coarse fines: settles quickly outside the work

area



Medium fines: can be transported large

distances during spring tide, prime case of

remote sedimentation



Fine fines: regularly transported large

distances, generally will not settle out,

contributing to suspended sediment impacts

Figure 11. Example sediment fall velocity distribution from Owen Tube test of fine material content of reclamation fill.

which is required to establish changes instress levels or partial mortalities of colonieslying off the transect line.

Example results from a repeat LIT surveyclose to the reclamation site at stationCR07 are presented in Table VI. The LITsurveys indicate no significant change inreef characteristics as illustrated by the plotin Figure 13.

For the exact repeat colony monitoring atthe same site, 50% of the colonies showedsome form of improvement in life formcharacteristics. 30% showed no changeand 2 colonies (20%) were noted to havedeclined as a result of physical damage notdirectly attributable to the reclamation works.

The sediment loading from the reclamationworks at this site over the monitoringperiod is tabulated in Table VII. Comparisonwith the coral tolerance limits presented in Table I and Table II indicates that thesediment loading falls in the No Impactcategory. This is consistent with therecoded LIT and exact repeat resultsconfirming, in this case, the applicability ofthe tolerance limits (at the No Impact level).Tolerance limits were therefore not updatedand spill budget limits for the period afterAugust 2006 were not adjusted.

Seagrass monitoringParameters used to assess the health of theseagrass areas include seagrass spatialdistribution and composition, seagrasspercent cover, seagrass diversity andevenness, seagrass biomass, sediment leveland composition.

Measures Analysis of Variance on Ranks,which is commonly used for comparisonbetween two datasets, is used for thestatistical analysis of sediment level andseagrass cover for comparison of the

baseline and repeat surveys. Figure 14 showsan example from a seagrass bed close to thereclamation site. The mean seagrass coverdocuments a general increase between thebaseline and the first Repeat Survey, but adecrease of approximately 20% documentedbetween the first and second repeat.

The corresponding sediment loading fromthe reclamation works at this site over themonitoring period is tabulated in Table VIII.This indicates the seagrass bed lie in theNo-Impact zone, though a moderate decrease

Figure 12. LIT Coral habitat survey in Singapore.

Table VI. Comparison of mean percent cover and standard deviationfor the major benthic categories at CR07

Baseline Repeat Survey 1 Repeat Survey 2Major Category August-05 May-06 August-06

Mean Cover (%) STDEV Mean Cover (%) STDEV Mean Cover (%) STDEVHard Coral 24.66 8.73 26.87 6.95 26.61 8.45Dead Coral 0.28 0.31 0.38 0.53 1.01 0.78Soft Coral 1.34 1.46 0.86 0.58 0.75 0.40Sponge 2.55 2.19 3.84 2.19 3.66 2.39Other Fauna 16.13 4.35 14.10 8.28 13.44 9.46Algae 19.93 8.01 27.87 8.76 30.36 12.86Rubble 32.72 16.21 21.86 8.62 22.02 9.51Rock 0.00 0.00 0.00 0.00 0.00 0.00Silt 0.00 0.00 1.94 2.50 1.63 2.16Sand 2.39 1.76 2.28 1.61 0.52 0.83Other 0.00 0.00 0.00 0.00 0.00 0.00



in cover was identified by the habitatmonitoring. In this case the hindcastmodels are conclusive in confirming thatthere is no direct flow of sediment from the reclamation area to this seagrass site,such that it can be firmly concluded thatthe decrease in seagrass cover is notattributable to the reclamation works.Tolerance limits were therefore not updatedand spill budget limits for the period afterAugust 2006 were not adjusted. The abilityto isolate impacts from a developmentproject from other third part or regionalimpacts is a major advantage of the feedbackEMMP system adopted in Singapore.

Sedimentation monitoring Sediment traps are deployed on the reefcrest, close to the LIT monitoring sites.These measurements documentsedimentation levels along the reef area,which is used in part to validate the resultsof the sediment plume hindcast models(incremental sedimentation abovebackground values) and in part to confirmtolerance limits. Sediment traps function asa measuring device for sedimentation onthe reef area and are deployed in threereplicates; each consisting of threecylindrical small tubes attached together.The theory and dimension of the sedimenttrap follows those recommended in theSurvey Manual for Tropical Marine Resources(English et al, 1997). See Figure 15 for animpression of the sediment traps deployed.

To function reliably in the high sedimentationenvironment present in Singapore, sedimenttraps are recovered every fortnight. As aresult of the large number of trapsdeployed in Singapore, ease of underwaterservice is important. This has lead DHI todevelop a single point of attachmentsystem that is operated by the single Allenscrew seen in Figure 15. This systemreduces the underwater service time byapproximately 50%, improves the reliabilityof the data by reducing sediment lossduring recovery and also reducesexpenditures associated with cable ties andother consumables by approximately 50%.

Table VII. Summary of percentage exceedence of suspended sediment and sedimentation loading over the coral reef monitoring site CR07 presented in Table VI

DateMarch April May June July August2006 2006 2006 2006 2006 2006

% Exceedence 5 mg/l < 5% < 5% < 5% < 5% < 5% < 5%Nett sedimentation kg/m2/day < 0.05 < 0.05 < 0.05 < 0.05 < 0.05 < 0.05

Figure 13. Changes in the mean percentage cover of

the major benthic categories.

Table VIII. Summary of sedimentation loading over the seagrass monitoringsites presented in Figure 14

DateMarch April May June July August2006 2006 2006 2006 2006 2006

Nett sedimentation kg/m2/day < 0.1 < 0.1 < 0.1 < 0.1 < 0.1 < 0.1

Baseline (Sep 05)

Repeat 1 (May 06)

Repeat 2 (Aug 06)

Figure 14. Comparison of mean seagrass cover along transect CY03.


Figure 16 presents an example of theabsolute sedimentation rates close to thework area at the same reef monitoringpresented in Table VI. This shows anaverage declining sedimentation ratebetween 0.08~0.11 kg/m2/day afterbaseline. The results presented in the figureindicate that no sedimentation impact atstation CR07 during July and August fallswithin the No Impact limits. These resultsare consistent with the results of thesediment plume hindcast and habitatsurveys (see Table VI for details of change inlive coral cover at CR07) and fall within theEQOs for the project.

Online turbidity sensors Online turbidity sensors are deployed at keyenvironmental receptors (coral reefs andintakes) in close proximity to the reclamationarea in order to provide an initial responsemechanism to any transients in suspendedsediment concentrations and to providesupplementary validation data for thesediment plume hindcast models. The instruments are vertically secured to aplatform deployed on the seabed, and heldapproximately 1 metre above the seabed.Data recorded is transformed from NTU toTSS via site-specific validation curves, whichare updated on a weekly basis based onmeasurements taken during instrumentservicing. The data is transmitted to a Data Information System that is used todisseminate all EMMP related data to theauthorities and contractors.

Figure 15. Three sedimentation traps are

fixed at each site located on the reef slope

close to the coral LIT sites. The height of the

trap from the reef surface to the opening is

35 cm. The sediment traps are held vertically

by angle-bars hammered deep into the

ground in an area of dead coral.

The actual dimension of the sediment trap is:

height 15 cm and Ø 5 cm.

Baseline

10 Jul ‘06

08 Aug ‘06

Average Sedimentation

CR07

Av

era

ge

Se

dim

en

tati

on

Ra

te [

kg

/m2/d

ay

]

Figure 16. Average sedimentation

rates at station CR07.

Figure 17. Left: deployed YSI turbidity sensor. Right: Time series of turbidity measurements.


Figure 17 presents a typical picture of theonline sensor and an example of meanturbidity levels. The increase in turbiditylevels observed in this example above thebaseline mean results from sensor fouling,which is a significant problem in Singaporewaters owing to high rates of algaegrowth, despite automatic sensor cleaningand weekly equipment service.

As the turbidity measurements provide onlya second level of EMMP response thereliability of the overall EMMP is notinfluenced by this fouling problem, whichwould otherwise be critical to managementplans reliant purely on static monitoring.

Other online instrumentation used forcontrol monitoring include, for example, noise meters (Figure 18) and AcousticDoppler Current Profilers (ADCP) (Figure 19).Noise meters are generally deployed atreceptor sites (residential buildings and/orwork sites) to document noise levels fromthe construction. ADCPs are deployed on the seabed for current and wavemeasurements.

CONCLUSION

The feedback approach to the EnvironmentalMonitoring and Management of reclamationworks summarised in Figure 20, which hasbeen adopted in Singapore, provides apractical and reliable method for the pro-active management of potentialenvironmental impacts resulting fromreclamation works.

The responsiveness of the system allowsunexpected impacts to be mitigated priorto them becoming a serious threat to theenvironment. Importantly, the level ofdocumentation provided ensures thatdevelopers and contractors are not exposedto unwarranted claims concerningenvironmental degradation as the EMMPapproach allows full segregation of projectimpacts from other third party disturbances.

In order to obtain the level of reliability andresponsiveness required to meet strict EQOsrelating to marine habitats and otherenvironmental receptors in Singapore, severalenhancements to various components of

the EMMP have had to be realised. Theseinclude empirical methods for estimation ofspill based upon sediment characteristicsand type of operation, adapting sedimentplume models to cater for complex dredgingand reclamation schedules, plus specificadjustment of settling and re-suspensioncharacteristics to cater for the complexitiesof reef morphology.

The performance of the feedback EMMP interms of meeting EQOs has been verifiedby habitat monitoring which also confirmadopted tolerance limits for corals andseagrass in high background suspendedsediment and sedimentation environmentssuch as those encountered in Singapore.

The EMMP techniques presented here havealso been successfully adopted for theenvironmental management of otherdredging and reclamation projects in theregion, including Bintulu and Kota Kinabalu,Malaysia and previously mentioned Bali TurtleIsland, Indonesia. The EMMP techniques arethus becoming accepted best practicemethodologies in the South East Asia Region.

Figure 18. A noise meter, built into a switch box. Figure 19. An Acoustic Doppler Current Profiler mounted on a stainless steel frame and about to be

deployed on the seabed.


REFERENCES

Duarte, C.M. & Chiscano, C.L. (1999) Seagrass

biomass and production: A reassessment.

Aquatic Botany, Vol. 65, 159-174.

Driscoll, A.M., Foster, T., Rand, P. and Tateishi, Y.

(1997). Environmental Modelling and

Management of Marine Construction Works in

Tropical Environments, 2nd ASIAN and Australian

Ports and Harbours Conference organised by

the Eastern Dredging Association, Vietnam.

English, S., Wilkinson, C. and Baker, V. (1997).

Survey Manual for Tropical Marine Resources

(2nd Edition), ASEAN-Australia Marine Science

Project: Living Coastal Resources. Australian

Institute of Marine Science, Townsville.

Harriott, V.J. and Fisk, D.A, (1988). Accelerated

regeneration of hard corals: a manual for

coral reef users and managers. Technical

memorandum GBRMPA-TM-16. Great Barrier

Reef Marine Park Authority, Townsville. 42 pp.

Hill, J. and C. Wilkinson (2004). Methods for

Ecological Monitoring of Coral Reefs. Australian

Institute of Marine Science, Townsville: 117 pp.

Møller J.S. (2000) Environmental Management of

the Oresund Bridge, Littoral 2000, Nice.

Thampanya, U., Vermaat, J.E., Terrados, J. (2002).

The effect of increasing sediment accretion on

the seedlings of three common Thai mangrove

species. Aquatic Botany 74, pp. 315-325.

Tomlinson, P.B. (1999) The botany of mangroves.

Cambridge University Press, United Kingdom.

Tun, K., Chou, L.M., Cabanban, A., Tuan, V.S.,

Reefs, Ph. Yeemin, Th., Suharsono, Sour, K., and

Lane, D. (2004). Chapter 9 Status of Coral Reefs,

Coral Reef Monitoring and Management in

Southeast Asia, 2004. In: Status of Coral Reefs

of the World, 2004. pp. 235-275.

Veron, J., Stafford-Smith, M. (2000) Corals of

the world, Volume I, II, III. Australian Institute

of Marine Science and CRR, QLD Pty. Ltd.

Waycott, M., McMahon, K., Mellors, J.,

Calladine, A., and Kleine, D. (2004). A guide to

Tropical Seagrasses of the Indo-West Pacific.

James Cook University, Townsville. 72 pp.

Figure 20. Summary of the prime components of feedback EMMP adopted in Singapore.

Documents

Environmental Monitoring and Management of Reclamations ... · ENVIRONMENTAL MONITORING AND MANAGEMENT OF RECLAMATIONS WORKS CLOSE TO SENSITIVE HABITATS STÉPHANIE M. DOORN-GROEN