lable at ScienceDirect

Ocean & Coastal Management 131 (2016) 25e38

Contents lists avai

Ocean & Coastal Management

journal homepage: www.elsevier .com/locate/ocecoaman

Challenges to implementing a ballast water remote monitoring system

Newton Narciso Pereira a, *, F�abio Belotti Colombo c, Marco Isaías Alayo Ch�avez c,Hernani Luiz Brinati b, Marcelo Nelson P�aez Carre~no c

a Department of Production Engineering, School of Industrial Engineering Metallurgical Volta Redonda (EEIMVR-UFF), Av. dos Trabalhadores 420, RoomC77- Vila Sta. Cecília, 27255-125, Volta Redonda, Rio de Janeiro, Brazilb Department of Naval Architecture and Oceanic Engineering, School of Engineering, University of S~ao Paulo, Av. Luciano Gualberto, 2230, 05508-030, S~aoPaulo, SP, Brazilc Department of Electronic System Engineering, School of Engineering, University of S~ao Paulo, Av. Luciano Gualberto, 2230, 05508-030, S~ao Paulo, SP, Brazil

a r t i c l e i n f o

Article history:Received 29 February 2016Received in revised form21 June 2016Accepted 19 July 2016

Keywords:Ballast waterMonitoringPortsWater qualityPollutionInvasive species

* Corresponding author.E-mail address: newtonpereira@id.uff.br (N.N. Per

http://dx.doi.org/10.1016/j.ocecoaman.2016.07.0080964-5691/© 2016 Elsevier Ltd. All rights reserved.

a b s t r a c t

In this paper, we describe a ballast water data logger system to monitor the ballast water exchange andthe water quality contained in ship tanks. This system is able to register physical-chemical parameters ofballast water by using sensors for measuring turbidity, salinity, dissolved oxygen, pH and temperature.Those data are tagged with the geographical position (GPS), date and time at which the ship operates itsballast system and are remotely transferred via satellite transmission to an Internet server. The systemwas installed on the ship M/V Crateus (from Norsul Navigation Company) and has been functioning sinceApril 2014, collecting ballast water quality parameters in the routes between Argentina and the northregion of Brazil. From the collected data, the system proved to be able to identify the ballast water ex-change along the ship's journey, allowing for independent verification of information provided by thecrew in the ballast water reporting form. As an additional advantage, this information can be auto-matically transmitted to the port authorities, improving the reliability of this information and reducing,or even removing, the possibility of data tampering. Nowadays, the salinity is the main indicator todeterminate whether a ship makes the ballast water exchange. However, we identify that water turbiditycan be one more indicator to identify the ballast water exchange that could be recommended by theInternational Maritime Organization.

© 2016 Elsevier Ltd. All rights reserved.

1. Introduction

The introduction of exotic species in port areas is currently agreat environmental problem in several parts of the world (Cohen,1998; Cohen and Foster, 2010; Pereira and Brinati, 2012), and thiswas first reported by the International Maritime Organization(IMO) in 1973 during the creation of the International Conventionfor the Prevention of Pollution from Ships - MARPOL 73/78 (IMO,2004; Cohen and Foster, 2010). In the course of the convention,Resolution 18 for Research into the Effects of Discharge of BallastWater containing Bacteria of Epidemic Diseases was approved,which charged the IMO with the responsibility for elaboratingmeasures of ballast water (BW) control (Cohen, 1998). In fact, since1994, several exotic species have been identified in many parts ofthe world (Hallegraeff, 1992; Carlton and Geller, 1993; Gollasch,

eira).

2006), and studies have identified BW as the vector of exotic spe-cies transfer (Ruiz et al., 1997). Therefore, the impact caused by theorganisms found in BW, such as Vibrio cholerae (Dobroski et al.,2009; Cohen and Dobbs, 2015), can be of great important for ma-rine environment, the economy and human health.

The first initiative taken by the IMO to deal with this problemwas to establish Resolution A.868 (20) in 1997, which recommendsthat ships perform ballast water exchange (BWE) in open sea. In2004, the International Convention for the Control and Manage-ment of Ships' Ballast Water and Sediments (BWM Convention)took place, with the purpose of establishing guidelines for BWcontrol (IMO, 2004).

BW operations normally occur while ships are unloading cargoin ports. In the port regions, the salinity varies between 32 ppt and35 ppt (parts per trillion), although it can be higher or lower insome specific cases (Murphy et al., 2008; Doblin et al., 2010; Cohenand Foster, 2010). On the other side, in open ocean, the salinityvaries, on average, from 35 ppt to 37 ppt (Murphy et al., 2008). Thus,

N.N. Pereira et al. / Ocean & Coastal Management 131 (2016) 25e3826

the BWE suggests that fresh water organisms cannot survive in saltwater and vice-versa (Smith et al., 1999). So, BWE typically elimi-nates between 70% and 99% of the organisms originally taken into atank while the vessel is in or near a port (Cohen, 1998). However, inorder to have a proof of the effectiveness of mid-ocean exchange,the BW salinity must be examined. This test consists of collecting asample of the BW in the tank, dripping it in a refractometer or usingan electronic meter and analyzing the salinity and specific weightof the sample. The result will confirm if the water collected origi-nates from estuary, coastal or mid-ocean waters and will confirmwhether this water was exchanged in the open s ea.

Current methods to verify the real exchange of BWat conditionsestablished by the BWM Convention are limited to determining thesalinity (Duggan et al., 2005; Drake et al., 2002; Choi et al., 2005),whilst it would be convenient if other parameters were also veri-fied. For an onboard verification, tanks have to be opened andsamples collected (Gray et al., 2007), but there are many reportedproblems concerning tank opening operations (Murphy et al.,2003), such as the lack of inspection stations, as well as the lackof specialized technicians to perform them. Another problem,although not yet reported in the literature, involves the teammobilization cost for this procedure. Alternatively, another methodfor this verification is to use the coordinates submitted in the BWreporting forms (BWRFs). From these reports, it is possible toidentify if the region of the BWE was at least 200 nautical milesaway from the coast and in waters at least 200 m (m) deep, thebasic demands for the exchange of BW (Pereira et al., 2014).

On the other hand, there is a problem associated with the reli-ability of the information in the BWRFs that the ships have todeliver to the Port State Control (PSC) before arriving at a port(Pereira et al., 2014). In fact, in the specific case of Brazil, records ofBWE violations are not uncommon. Leal Neto (2007) presented themain problems found in a survey conducted by the GloballastProgram, using forms delivered to the Brazilian Navy between 2001and 2002. Caron Junior (2007) showed inconsistencies during theanalyses of 808 BW forms handed to maritime authorities of theport of Itajaí-SC, in the south of Brazil. The Brazilian HealthSurveillance Agency (2003) conducted another study that showsthe results of 99 samples of BW in 9 Brazilian ports, revealing thatsome ships had not exchanged the BW. However, the lack of con-fidence of BWE also affects other countries. In fact, Brown (2012)showed the non-compliance BW report in California. In the firstsemester of 2012, approximately 1 million metric tons (MMT) non-compliance BW was discharged in California ports, due to eitheroperational error or incorrect geography and not to intentionalmismanagement. In 2014, all ships that accessed the Great Lakeshad their tanks examined and BW samples were collected (GreatLakes Seaway, 2015). Pereira et al. (2014) presented several prob-lems with BWRFs delivered by ships to the Brazilian Navy in theAmazon region, where it was identified that ships deballast in portsin this region without conducting the BWE at sea. Additionally, itwas identified that these ships presented problemswith the qualityof the BW inside their tanks.

Generally speaking, the BWRF does not provide informationabout the quality of the BW captured by the ships. However, in-formation of water characteristics, such as turbidity, salinity, dis-solved oxygen (DO), pH and temperature, would be very useful forthe PSC to identify the quality of the water, especially because thewater collected in the ports may contain domestic, industrial andagricultural effluents (Vandermeulen, 1996; Peterlin et al., 2005).These water parameters can indicate the probability of survivingspecies and treatment efficiency of the BW inside ship tanks. In fact,some of these effluentsmay be highly polluting when discharged innature in the destination port environment. Among these constit-uents of estuary waters can be found dissolved solids, salts, organic

sewage, nutrients, heavy metals, hydrocarbons, radioactive mate-rials and herbicides (Clark, 1986) that are not identified with cur-rent methods utilized to evaluate the quality of BW. Mainly, therelease of sewage into port regions can alter the pH and the de-mand for DO in water, carrying the nutrients and promoting theproliferation of toxic algae and the destabilization of the aquaticecosystem (Morrison et al., 2001).

In environments with high concentrations of organic matter,such as algae, turbidity can be changed (Torgan, 2011). Theseorganic matters can be transferred into the BW tanks and can causeproblems such as red tide when discharged in other environments.The change in turbidity can also occur in the presence of solid“sand” in suspension (Prange and Pereira, 2013). The presence ofdissolved solids in the seawater tends to be higher than in estuariesdue to the low sea depth. Therefore, it is possible to find a largeamount of sand at the bottom of BW tanks (Prange and Pereira,2013). Thus, the turbidity is indicative of the presence of severalorganic and inorganic components that may be present in the BWcaptured by the ship.

Since salinity may change from port to port, monitoring thisparameter may indicate whether there are risks of transfer speciesdue to the similarity between origin and destination ports. Otherfactors associated with environmental similarity are water pH andtemperature, which can significantly impact the survival and sta-bility of toxins produced by many organisms (Torgan, 2011). Thus,the variation in the pH of the water collected and the waterexchanged by the ship can be an indicative of the probability ofsurviving species in the ballast tanks. Besides, the presence ofdissolved gases in the BW can also indicate the probability of sur-viving organisms and is related to the water temperature (Jewettet al., 2005). For example, there are certain ranges of dissolvedoxygen (DO) inwater (mg/l) that can be translated into survivabilityof species. So, low DO levels are responsible for the death of manyorganisms in BW (Tamburri et al., 2002).

As can be seen, in spite of the importance of different waterquality parameters to understand the effectiveness of BWE process,only the salinity is evaluated when ships arrive at the port andundergo BW inspection, with no other parameters reported byships in BWRFs. So, motivated by this, in this paper, we develop aBW remote monitoring system (BWRMS) that collects, in real timeand directly inside the ballast tanks, the BW quality parameters.This system includes sensors for turbidity, conductivity (salinity),dissolved oxygen (DO), pH and temperature. An important char-acteristic of this monitoring system is that the data from the sen-sors are tagged with the ship's geographical position and the dateand time of the collection. The data are recorded in an unchange-able electronic controller unit and remotely transferred via satelliteto an inland web server, where they can be analyzed. In that way,the system allows not just monitoring and analyzation of theevolution of water quality parameters but also independent veri-fication of the geographical position where the BW is exchanged.This can be done from the variations observed in the data collectedfrom sensors. Note that since the collected data are automaticallytransmitted, the relevant information can be directly sent to theport authority to validate the information in the BWRF, improvingits reliability and reducing or even removing the possibility of datatampering. Even more, this can be the starting point for a new typeof electronic BW reporting form (e-BWRF).

This system was installed on the ship M/V Norsul Crateus (fromNorsul navigation company of Brazil) and has been in function sinceApril 2014, collecting BW parameters in the routes between theports located in South America (Argentina) and Brazil. In this paper,we will present the validation of the system considering the datafrom all voyages realized by the ship during the time between April2014 and December 2015, where it is possible to compare the data

N.N. Pereira et al. / Ocean & Coastal Management 131 (2016) 25e38 27

from BWRMS and the BWRF sent by the ship. We selected onespecific voyage to show a zoom the data collected from the BWRMSand the BWRF from M/V Norsul Crateus. This voyage occurred fromPort of San Nicolas in Argentina (at Paran�a river) to Port of Santosduring the time between December 2014 and January 2015.

2. Materials and methods

2.1. System development

The development of the BWRMS was aimed at allowing for themultipurpose use of the system. Therefore, the system should beflexible and mostly adaptable mainly to the ship ballast operatingconditions. Fig. 1 presents the conceptual model of the system.

As can be seen in Fig. 1, the central component of the system isthe control unit, which collects data from the sensors, correlatesthem with the vessel's geographical position (from GPS) at whichthese data are collected and conditions all information to transmitto an inland server via satellite communication. Since there is not acommercial solution to attend to these requirements (collectionand transmission of sensors plus GPS data), a dedicated electroniccontrol circuit was designed and mounted.

The system includes sensors to measure conductivity, turbidity,temperature, pH and dissolved oxygen, which were purchasedfrom Global Water Instrumentation. The sensors were selected tomonitor physical-chemical parameters and to perform measure-ments in fresh and brackish water due to characteristics of fluvialbasins in Brazil. However, organic and biological parameters canalso be measured and sensors acquired from any supplier companycan be utilized (next step). The characteristics of the sensors arepresented in Table 1.

The vessel's geographical position coordinates are obtained bymeans of a Global Positioning System (GPS) from GARMIN (modelGPS 17x NMEA 0183). To guarantee an autonomous functioning, asolar energy-based power supply system was incorporated, ofwhich a proposal guaranteed the system's operation to be inde-pendent of the ship's own power sources. For this purpose, a

Fig. 1. Conceptual model of the system

photovoltaic solar panel captures sunlight and charges a 12 V bat-tery, which, in turn, provides power for control units. This schemepermits the system to operate without any crew interference andpreserves its integrity.

The system transmits the collected data using a Digi m10 mo-dem and Orbcomm satellite constellation. The data are sent in bi-nary format through messages to the available communicationsatellite, which retransmits them to an inland server. Software wasdeveloped to collect the data from the server and store them in adatabase, from where they are accessible (via Internet) to the enduser. For this, we also developed a web-based graphic interface forinterpreting the numbers. In the results reported here, the collecteddata are transmitted every 55 min, which seems suitable for thistype of study. However, this time is determined by the contracteddata transmission service and can be significantly shorter.

2.2. Ship installation process

The system was installed on the dry bulk carrier M/V NorsulCrateus, IMO Number 9056399, from Norsul Navigation Company,operating in Brazilian cabotage navigation. It has 42,487 DWTand aBW capacity of 26,710 m3. This ship also operates in the Amazonbasin and the River Plate basin, using fresh, brackish and salt waterduring the ballast operation.

The BWRMS was installed on April 9, 2014, in the Port of Santos(Latitude: 23�87019.1200S, Longitude: 46�37081.9700W) during theunloading of iron ore cargo at a private terminal. The controlelectronic system (control unit, data acquisition board and battery)was placed inside a dust-proof, water-resistant case and installed inthe boatswain locker at the bow castle of the ship.

The sensors were installed in superior portside tank 1 A (ca-pacity of ±600 m3), one of the 14 ballast tanks of the ship. Forprotection and mechanical support, the sensors were installed in astainless steel cage welded to the inside tank wall. This tank is closeto the forecastle and was selected due to the facility of sensorinstallation (the tank is just below the deck), the proximity to theboatswain's locker (the cable length is limited to around 30 m) and

installed in M/V Norsul Crateus.

Table 1Characteristics of the sensors used to monitor BW quality.

Type of sensor Model Sensor range Sensor output Accuracy Power required

Conductivity WQ301 0e42000uS 4-19 mA 1% of full scale 10-30VdcTemperature WQ101 �50e50 �C 4-19 mA ±0.1 �C 10-36VdcpH WQ201 0e14 pH 4-19 mA 2% of full scale 10-30VdcTurbidity WQ730 0e50 0e1000 NTU 4-20 mA ±1% full scale 10-36VdcDissolved oxygen WQ401 0-100% saturation 4-19 mA ±5% full scale 10-36Vdc

N.N. Pereira et al. / Ocean & Coastal Management 131 (2016) 25e3828

the fact that it performs BWE procedures every voyage.To install the sensors, it was necessary to open small holes

(10 cm in diameter) on the main deck of the ship and on the wall ofthe boatswain's locker of boatswain wall. Sets of stainless steelflanges were welded on these accesses to pass through the sensorcables and isolate the tank. These flanges include water feed-through to guarantee the perfect impermeability of the BW tankand the wall of the boatswain's locker.

Finally, the antenna, GPS, modem and solar panel were installedat the top of the foremast, the base of which is just above theboatswain's locker and the control system case, at a suitable dis-tance to extend the data interconnection cables. During theinstallation process, the stainless steel components were welded bythe ship's crew to fix the different parts of the system (sensor cage

Fig. 2. General view of the BW monitoring syste

in the BW tank, flanges on the main deck and the wall of theboatswain's locker and solar panel at the foremast). A general viewof the installation of the BW monitoring system is shown in Figs. 2and 3.

Fig. 3 shows some installation details of different parts of thesystem. The entire installation process took around 16 h to beconcluded (from the initial presentation to the ship crew up to thefinal turning on of the system) and directly involved 6 people.

2.3. Data evaluation

The data collected from the sensors are stored in a databaseaccessible via the Internet. In that way, the collected data can beanalyzed through a webpage that allows for selection of the period

m installed on the M/V Norsul Crateus ship.

Fig. 3. Details of the BWRMS installation in the ship: (A) cage of sensors inside the tank, (B) stainless steel flanges to pass through the cables of sensors, (C) system of protection ofcables, (D) control system installed in the boatswain's locker and (E) installation of antenna, modem and solar panel in the foremast.

N.N. Pereira et al. / Ocean & Coastal Management 131 (2016) 25e38 29

of time of interest and shows (over a map) the route of the shipduring this period. Small icon pointers on the route indicate theexact position where the data are collected, and a dialog box shows(in each pointer) the values measured by each sensor. Beside it,graph windows show the variation of each parameter during theentire requested period. A typical view of this web interface isshown in Fig. 4.

Although it can be interesting as a first printout to observe longperiods of time, the interface shown in Fig. 4 is not very friendly forthat. In fact, due to the large amount of available data, the time ofthe queue and port operation periods, the data observations can besomewhat confusing. For these cases, the system offers the possi-bility of exporting a data file (in Microsoft Excel format) for furtheranalysis.

3. Results

We present results from this development, reporting the stepsof the implementation and validation of the BW monitoring sys-tem. Challenges are presented, showing how problems were solvedand the learning curve of this investigation. In the literature, a formof BW monitoring with these characteristics was not yet reported,nor was there this level of effort to build a specific and dedicatedsolution for BWE compliance.

3.1. Initial measurements and considerations

At this moment, the BWRMS developed is continuously moni-toring the tank parameters and has been since its installation inApril 2014, collecting data during the voyages for the BWE pro-cedures and even when the ship is stationary at the port or in highseas.

Although the system had been installed in April/2014, untilDecember/2014, we were just testing and adjusting the systemfunctioning. During this period, we experienced intermittentproblems with three sensors (temperature and pH) and with thedata transmission. In fact, after an in loco check-in, it was clear thatthese problems were related to defective sensors and to failures inthe power delivery to the system.

The failure in the power supply was related to accumulation ofan opaque layer of sea salt aerosol and dust from ship iron oreoperation over the solar panel, which prevented efficient solar-to-electrical energy conversion. Consequently, the battery-recharging process was ineffective and, after a few days, the bat-tery lost its charge and the system turned off. In the sequence, thebattery recovered a partial charge and the system turned on again,in a cycle that repeated continuously and explained the observedintermittent failures in data collection and transmission. Thisoccurred between May and October 2014.

Fig. 4. Typical view of the web user interface developed to analyze the data collected from the BWmonitoring system. The users can make a query in function of the specific period.Data are sent to the server in 1-h intervals and are plotted in the interface, showing the ship's route and the geographic coordinator with graphics of the sensors' readings.

N.N. Pereira et al. / Ocean & Coastal Management 131 (2016) 25e3830

To overcome this problem, we could perform periodic cleaningof the panel or even replace the panel with a bigger onewith higherpower capacity. However, due to logistic and economic reasons, thiswas not possible, and a simpler solution was to connect the systemto the main power energy of the ship. Note that this does notrepresent a limitation of our system, since the photovoltaic panelcan be resized or installed in a cleaner place or a cleaner ship. Evenmore, actually, the photovoltaic panel is actually not essential to theBWRMS operation, and it was included just to guarantee a relativeautonomous operation to the ship crew. Regarding the sensors, wehad used them in the lab before installation on the ship. The pHsensor worked only 8 days inside of the ship tank after the instal-lation on board. The temperature sensor operated for 138 days. Thisshows that the marine environment and the aggressive conditionsof the BW ship tank can make monitoring using water qualitysensors difficult.

3.2. System validation

After the energy problemwas solved, the data from the BW shiptankweremore accurate. Therefore, we started to compare the databetween the BWRMS and the BWRF. The data from all the voyagesof the ship were plotted during the period from April 2014 untilDecember 2015. First, the data from ship voyages acquired from theBWRMSwere plotted on a ship timeline (blue line) in the horizontalaxis. Due to the decreases in pH and temperature, only the datafrom sensors that were working during that time were plotted.

Each event relative to the ballast (red line), BWE (blue line) and

deballast (green line) operations were plotted in the vertical axis,considering the date and time of the event reported in the BWRF.From May to October, the system was unstable and data were lost,and it was not possible to correlate that from the BWRMS and theBWRF. Because of that, there is no vertical line that represents theBW operation from the ship. We divided the period of analyses intwo figures. Fig. 5 presents the results collected during the first yearof system operation, from April 2014 to April 2015.

We could note that the BWRMS detected variations in BWquality during the ship's voyages in the first year. All vertical lineswith a register of BWoperation by the ship were identified from theBWRF's. After the BWRMS installation, the BW tank was ballastedwith water from the Port of Santos, and the system started theoperation. During the time period between April 2014 untilDecember 2014, this ship ballasted only with sea water(conductivity ¼ 42,000, mS ¼ 30.2 psu in 20 �C).

On 25 December 2014, a ship ballasted in the Port of San Nicolas(PSN) (located in the Paraguay-Paran�a waterway - Latitude:33�35009.5600S, Longitude: 60�17054.1500W) in Argentina to go to thePort of Santos (PS). Because of that, the conductivity reduced from42,000 mS to 15,000 mS (9.8 psu), and the turbidity changed from 72NTU to 212 NTU, that being a characteristic of fluvial rivers.

The ship arrived in PS on 01/05/2015 and deballasted, and bal-lasting again at PS. The BWRMS detected a variation of BW qualityon 01/08/2015, but the BWRF showed that BWE had occurred on01/07/2015. The BWE occurred near of Montevideo city, in Uruguay(Latitude: 34�98068.0500S - Longitude: 54�57054.4700W). At thispoint, the salt water salinity is influenced by the flow of fresh water

Fig. 5. Register of ship voyages and BWoperation, from April 2014 to April 2015. The horizontal line (blue) represents the sensors' reading. Vertical lines represent the BWRF events(red -ballast, blue - BWE and green - deballast). Some ballast events occurred on the same date where the ship deballasted/ballasted in the same point. Because of that, red andgreen points appear nearby with a time delay. Each BWE event is numbered above in sequence of occurrence (blue numbers). (For interpretation of the references to colour in thisfigure legend, the reader is referred to the web version of this article.)

N.N. Pereira et al. / Ocean & Coastal Management 131 (2016) 25e38 31

from La Plata Basin, which reduces water salinity, making itbrackish water (Piola et al., 2005).

Therefore, the conductivity sensor showed a reduction in similarreadings in the last voyage and an increase in turbidity. In this case,the ship deballasted in PSN. The ship ballasted again in PS on 01/28/2015, where it is possible to note a variation of water turbidity inrelation to previous water in the tank. Until 03/19/2015, all BWRFsshowed that the ship ballasted in the Port of Santos, BWE in mid-ocean and deballasted in PSN, but the conductivity value was sta-ble in same value read where it could indicate eventual sensorreading problems. However, the turbidity and DO showed varia-tions correlated with BW events detailed in the BWRF.

After 03/19/2015, the ship moved to the Ponta da Madeira Ter-minal (TPM), in Maranh~ao State (Latitude: 02�56065.0900S, Longi-tude: 44�41015.1400W), and realized the BWE at 03/22/2015 nearMaca�e City in Rio de Janeiro State (Latitude: 22�03090.0000S,Longitude: 40�41030.0000W). The ship deballasted in TPM andmoved to Port of Usiba Terminal (TPU) in Bahia State (Latitude:12�82057.5700S, Longitude: 38�49088.8200W), where it ballasted andmoved to Shipyard Enavi in Rio de Janeiro (Latitude: 22�85088.5600S,Longitude: 43�10059.0100W) to dock (Fig. 6).

According to the BW registers of the first year, we could notethat during the ship's voyage on the sea, the BWEwas realized closeto the Brazilian coast, as shown in Table 2. These data were collectby the BWRF and compared with the BWRMS to identify whetherBW quality suffered alternations in these positions. Distances werecalculated with Google Earth® Fig. 5.

In the second year, the ship made voyages around the Braziliancoast, concentrated in the northeast region. The ship was dockedbetween May and June. During this period, the readings from theconductivity sensor showed variation while the ship was at theshipyard. It was impossible to determinate the real causes of thiseffect. The temporal series of the second year is presented in Fig. 7.

Over the period of 07/01/2015 until 10/23/2015, all ship ballastoperations occurred in the Amazon region. The ship operated be-tween the Port of Alumar (PA) in Maranh~ao State (confluence be-tween Estreito dos Coqueiros and the River of Cachorros near theS~ao Marcos Basin e Latitude: 02�67’.79.3800S, Longitude:44�36010.2000W) and the Port of Trombetas (PT) in Par�a State (inTrombetas River - Latitude: 01�45.5006800S, Longitude:56�39.8907500W), and during all voyages, the ballast occurred in PA.

Only on 08/23/2015 did the ship ballast in PA and deballast inthe Port of Juriti (PJ) in Par�a State (Latitude: 02�17044.7600S -Longitude: 56�11030.6000W). On 10/23/2015, the ship ballasted in PAand moved to TPM, where it deballasted on 10/28/2015. As the shipwas in the same region the BWE was not performed.

The last ballast operation occurred at TPU, on 11/13/2015, whenthe ship was ballasted with salt water, but the conductivity sensordid not identify this variation in the salt water. In this period, allBWE operations usually occurred when the ship was entering theAmazon river, which has a low salinity and high turbidity.

On the other hand, we could identify that similar turbidityvalues were identified by Alcantara et al. (2010); Affonso et al.(2011) considering the measurements in situ during all seasonsthroughout the year in Amazon rivers. DO measures showed valueshigher than the first year of monitoring. This is a characteristic ofAmazon rivers, and similar DO values were measured by Quay et al.(1995) in terms of DO saturation. These values suggested that dueto the Amazon water characteristics, the oxygen utilization ratescontinue inside of the tank for some period of time during theship's voyage, as a cycle process that is always activated when a BWoperation occurs. This effect was also observed in Johengen et al.(2007) while the BW tanks were monitored and several registersof BW reoxygenation inside of the tanks occurred. During the firstyear, the ship did not ballast in Amazon rivers, and the DO level islower than in August and October 2015.

Fig. 6. Position of BWE's informed by BWRF detected by BWMRS. Brazil map shows the region where the ship was operation during these voyages.

Table 2First year observations of BWE position by the BWRFs considering the sequence of voyages in a temporal series for ballast event (red line).

Voyage Date Coordinate Local Distance of coastnautical miles

1 4/14/2014

Latitude:32�02’.61.6000S Littoral of Rio Grande do Sul e South Atlantic Ocean 48.22Longitude:50�58083.8000W

2 11/14/2014

Latitude:28�39’.0000S Littoral of Santa Catarina e South Atlantic Ocean 13.92Longitude:48�40'.0000W

3 12/13/2014

Latitude:28�34054.0000S Littoral of Santa Catarina e South Atlantic Ocean 9.8Longitude:48�37024.0000W

4 1/7/2015

Latitude:29�43080.0000S Littoral of Rio Grande do Sul e South Atlantic Ocean, but the BWRMS detected water quality variation atLatitude: 34�98068.0500S, Longitude: 54�57054.4700W in 1/08/2015

31.34Longitude:49�23020.0000W

5 2/1/2015

Latitude:26�42039.4000S Littoral of Santa Catarina e South Atlantic Ocean 52.14Longitude:47�42091.6000W

6 2/18/2015

Latitude:24�57060.0000S Littoral of Sao Paulo e South Atlantic Ocean 56.22Longitude:46�46070.0000W

7 3/22/2015

Latitude:22�03090.0000S Littoral of Maca�e City in Rio de Janeiro State 22.89Longitude:40�41030.0000W

N.N. Pereira et al. / Ocean & Coastal Management 131 (2016) 25e3832

In the northeast of Brazil, it is warmer than in the southernregion, where the ship had made most of its voyages during 2014.Because of that, the DO could be influenced by the temperatureinside of the ship's tank. Johengen et al. (2007) observed the in-fluence of low temperatures on DO measurements inside of BWtanks.

Our measurements presented that data from turbidity and DOwere adherent with the BWRF operations reported during thecourse of the investigation. When the BWRF information wascompared with the information from the BWRMS, sometimes asmall variation occurred in the times reported in the BRWFs and theBWRMS. This occurred because the ship crew could register the BWoperation time differently than was detected by the system or itcould also be due to some delay of data transfer.

To identify where the BWE's occurred, Fig. 8 and Table 3 presentthe analysis of the BWE positions realized by the ship in the second

year informed by the BWRF.Table 3 shows that the interval of the BWE operation was

reduced because the ship made short voyages during this period,concentrated in the same region.

To evaluate the system validation by one specific voyage, Fig. 9presents details of BW operations detected by BWRMS during avoyage from the Port of San Nicolas to the Port of Santos between12/25/2014 and 01/05/2015 and compared the accuracy of thesystem to the BWRF information.

Fig. 9 shows that the ship ballasted in PSN and deballasted in PS.The BWRF information was coherent with the sensors reading andpresented evidence that BWE had occurred in this voyage. Thisexample demonstrated that it is possible to verify the BW qualityinside of the tank before the ship's mooring. Additionally, the BWRFinformation can be verified and correlated with the place, time anddate of the BWE reported. In this voyage, the ship did not realize

Fig. 7. Register of ship voyages and BW operation from April 2015 to December 2015. The horizontal line (blue) represents the sensors reading. Vertical lines represent the BWRFevents (red e ballast, blue e BWE and green e deballast). Each BWE event is numbered above in sequence of occurrence (blue numbers). (For interpretation of the references tocolour in this figure legend, the reader is referred to the web version of this article.)

Fig. 8. Position of BWE's informed by BWRF detected by BWMRS. Brazil map shows the region where the ship was operation during these voyages.

N.N. Pereira et al. / Ocean & Coastal Management 131 (2016) 25e38 33

Table 3Second year observations of the BWE positions.

Voyage Date Coordinate Local Distance of coast nautical miles

8 7/17/2015 Latitude: 01�60 95.0000N In Par�a State near the mouth of the Amazon river e northeast Atlantic Ocean 80.91Longitude: 48�31091.6000W

9 7/28/2015 Latitude: 01�01050.0000N In Amap�a State near the mouth of the Amazon river e northeast Atlantic Ocean 78.4Longitude: 48�25030.0000W

10 8/20/2015 Latitude: 00�5300.0000N In Amap�a State near the mouth of the Amazon river e northeast Atlantic Ocean 7,84Longitude: 49�51047.2000W

11 8/31/2015 Latitude: 01�10.9000000N In Amap�a State near the mouth of the Amazon river e northeast Atlantic Ocean 46Longitude: 49�13’.80.0000W

12 9/14/2015 Latitude: 00�5028.4700N In Amap�a State in the Amazon river at Macap�a city e northeast region 0Longitude: 50�57’.70.0000W

13 9/29/2015 Latitude: 01�1102.2000N In Amap�a State near the mouth of the Amazon river e northeast Atlantic Ocean 38,2Longitude: 49�19’.75.0000W

14 10/11/2015 Latitude: 01�08030.0000N In Amap�a State near the mouth of the Amazon river e northeast Atlantic Ocean 21,75Longitude: 49�31”.80.0000W

N.N. Pereira et al. / Ocean & Coastal Management 131 (2016) 25e3834

that BWE and brackish water was discharged in the estuary ofSantos. This happened because the ship was transporting brackishwater to be discharged in a salt water environment. Then, itgenerated an osmotic stress, where the fresh and brackish speciescannot survive in salt water.

On the other hand, the conductivity sensor reading presented adelay in identifying the change of BW quality. This suggests thateither this sensor could be influenced due to the latest operationonly being in salt water or the sensor quality not being able todetect high variations of conductivity in a short time. We verifiedsimilar events in other voyages in the northwest region during thesecond year.

However, the turbidity sensor reading changed at the samepoint indicated by the BWRF. The turbidity values for the LowerParan�a River near San Nicolas are about 30e230 NTU (Vazquezet al., 1998), and for this voyage, the sensor read similar values.The DO sensor reading concentration rates in the San Nicolas regionwere similar to those identified by Casco et al. (2014), and it de-pends on the year's seasons and temperature influence.

The variation of the sensors reading could be due to the acrylicbox utilized to keep the sensors wet all the time. Because of that, itwas impossible to verify the period in which the BW tank wasempty. This box can accumulate more organic matter and influencethe behavior of the sensors' reading, but it shows that there is anorganic dynamic inside of the tank during the ship's voyage.

When the tank was empty, sensors were reading only the pa-rameters inside of the acrylic box from the cage of sensors.Another aspect is that when the BW tank is empty, the liquidaccumulated inside of the box tends to be in motion during theship voyage. In this condition, there is reoxygenation and gasesinside of the tank that are in contact with liquid and sedimentsaccumulated on the bottom of the box that can influence thesensor readings.

4. Discussion

We have reached the main objective of this study of thedevelopment of a system able to validate the BWRF informationcompliance. It then shows that it is possible to monitor the BWquality inside of BW tanks and correlate correctly the real positionwhere the BWEwas performed by the ships. This is an efficient wayto avoid occurrences that were reported in various studies (JuniorCaron, 2007; Leal Neto, 2007; Brown, 2012; Pereira et al., 2014)and control the quality of BW discharge in sea and fluvial ports.

Considering that a BW treatment system that reaches the Cali-fornia standard does not exist (Commission, 2014) and they did nottest the real condition to eliminate pathogenic species (Cohen and

Dobbs, 2015), the BWE then continues being an efficient alternativeto eliminate invasive species in BW tanks while a universal stan-dard is not defined.

For this reason, BWE monitoring and BWRF informationcompliance are ways to guarantee that ships are realizing the BWEin the correct geographic position and discharging water followingthe IMO standard. With this system, this verification can be doneduring the entire time of the voyage of the ship, and it can becompared with the BWRF information sent by ships to the Mari-time Authority. It is an independent form to assist the PSC andenvironmental agencies in monitoring the quality of the BW dis-charged in their ports and making decisions about the acceptanceof the ships in the port. During algae bloom and BW operation atnight, it is possible to identify whether the BW tank is loadingwater with high turbidity at sea.

However, installing and controlling a BWRMS is not simple, andchallenges are present in the installation of sensors inside of BWtanks, controlling the sensor durability, system maintenance, thequality of data gathered and the validation of BWRF information.

Similar challenges were reported in previous attempts tomonitor BW presented by Johengen et al. (2005, 2007). Some of theproblems and the results in this study can be compared in terms ofBW quality monitoring inside of BW tanks of the ships Irma andM/V Lady Hamilton.

However, the proposal of both studies is the monitoring of BWtanks, while themain difference between these studies refers to thedevelopment of a particular solution for the remote BW moni-toring. In these previous studies, the authors had installed a con-ventional quality water instrument to evaluate water parameters.

This application did not allow that sensor readings were ob-tained in real time. Johengen et al. (2005, 2007) also presented thatthe GPS data collected had poor quality and could be used inconjunction with water measurements by instruments. Then, datafrom master's ships that reported the BW operations were thenused to compare with the data collected by sensors with BWE. Inour case, we could identify the position where BWE occurred onlyby observing the BW quality alteration in the web interface, whichshows the ship's geographic position with water quality graphicsfor each parameter monitored.

The process of BW monitoring can be done without anyparticipation of the ship's crew. In this previous case, the authorsreported that whenever the ship's crew was solicited to monitorthe water quality using some instrument, this operation failed. Inone event, the equipment was damaged, and this shows that utilityof the BW remote monitor. In our case, the ship crew was involvedonly in the system installation. After that, the system operationoccurred independently, without the ship's crew.

Fig. 9. Register of ship voyages and BW operation from 12/25/2014 to 01/05/2015.

N.N. Pereira et al. / Ocean & Coastal Management 131 (2016) 25e38 35

The other aspect is that the water quality instrument wasinstalled directly in the tank and did not suffer any influence fromthe acrylic box inside of the cage of sensors. However, the obser-vations were similar in terms of high turbidity of BWwhen the shipwas operating in river waters and low in sea water.

In these cases, ships that have been ballasting in rivers presentmore probability of accumulating sediments inside of tanks. In this

way, our acrylic box was very sensitive to collected sediments, andwe observed this during the ship's visit to sensors' maintenance. Inregards to the conductive sensors, the range of the scale was higherand allowed for monitoring of high values in terms of BW salinity(ppt) than this study using the bottom of the scale in salt water.During the first year of monitoring, the ship ballasted more in seawater than in fresh water. It is obvious that it was associated to the

N.N. Pereira et al. / Ocean & Coastal Management 131 (2016) 25e3836

ship market's demand for transport.This ship has been operating only in the cabotage routes in

Brazil, and the ballasting/deballasting always occurred close to thecoast. In contrast to ships that operate in marine navigation,cabotage ships will usually deballast near the coast and regionswhere the BW can be mixed with brackish and fresh water.

The monitoring of the ship's deballast position from cabotagenavigation can assist the PSC in determining the better region toconduct this operation. The decision to realize the BWE is frommaster ships that will evaluate where they can start this operation.Our measure calls attention to this operation on the Brazilian coastand probably has to be the same in other countries. It is associatedwith operation cost, distance between call ports and time to com-plete a travel. Then, it is natural that the cabotage ship will sail nearthe coast region and will conduct the BWE operation in this region.For example, in the Amazon region, the river can reach more than200 km into the sea, and water will be mixed with sea water.Because of that, the remote monitoring of BWE is important andcan map the more common area of ship BW discharge during theexchange operation. Considering this characteristic and full-timemonitoring of BW quality inside of the tank for a long period oftime, the results were not examined in previous studies.

On the other hand, considering the period during which wehave been monitoring the BW tank, it is in fact a very aggressiveenvironment with some air and illumination limitations.

This condition has a great influence on sensors' durability andgenerates a difficulty of maintenance. It could be confirmed due tothe short time of operation of the temperature and pH sensorsinside of the BW tank.

It is then necessary to consider that a ship completes manytravels during the year in different ports and routes. To conduct theBWRMS repair, it took three visits in total to detect problems withthe solar panel and solve the energy failure to correct the operationof the system. During this period, sensors failed, and due to the lackof data transfers, we could not identify the cause of the sensors'reading problems. This showed that the sensors'maintenance couldbe a great challenge for correcting the operation of this system.

Additionally, we believe that the closed cage of sensors canaffect the results of the sensors' readings, considering that somesediments, salt, organic matter and other compounds can bedeposited in the bottom. The sensors' quality can also affect thedata gathered and the satellite service. We tested only one supplierof sensors, and other types and suppliers can be used in thisoperation. For example, Raid et al. (2007) utilized a multi-parameter instrument probe from YSI 6600EDS but opted to useseparate sensors and to construct our monitoring solution appliedspecifically to BW tanks. Another challenge to the BWRMS opera-tion is sensor calibration, which has to be done frequently andsubstituted when necessary. It has to obligate the manager of thisoperation to create a program of inspection and cleaning for thesensors.

Another point that can influence the sensors' reading is theship's motion and external environment's conditions. Our resultsshow some alterations during the temporal series of water qualitythat could be altered due to the acrylic box during the ship's motionin bad weather conditions. One solution to solve this problem willbe remove sensors of box and inserting an accelerometer sensor toread the ship's motion and correlate it with the BW quality. Thetemperature sensor mainly represents the external environment'sinfluence above the BW tank during the ship's voyage. In the courseof the first measure with this sensor operation, we were able tonote some alteration in pH and DO levels with tank temperaturevariation.

Even with these challenges presented, the BWRMS demon-strated that it is possible to monitor the BW quality remotely, and

this can be an advantage of BWM when compared to currentmethods to evaluate the BWRF compliance. Moreover, this BWremote monitoring can assist us in understanding the BW qualitydynamic inside the tank. When the monitoring is punctual, thetemporal series of quality variation inside of the tank can be lost. Tounderstand this behavior, it is necessary to monitor the tankconstantly.

Associated with this, during the ship's voyages, data collected inseveral ports can create a port water-quality database to validatethe BWE operation in coastal ports.

Thus, the system was developed to allow for the coupling ofeight sensors, but this can be expanded so that a single control unitmay monitor all the ballast tanks of a ship through the addition ofsensors and A/D converter units. Themodem, GPS and satellite datatransfer systems are located in an external area of the ship. There isno need for multiple control units if one means to measure datafrom all the ballast tanks. In this case, it is better to place the A/Dconverter units close to the ballast tanks. Thus, a single pair of ca-bles can run along the ship to transmit data from all ballast tanks tothe control unit. The prototype includes a data acquisition unit inthe hermetic booth, but this is not an essential requirement; it wasused for simplicity purposes.

The equipment for capturing solar power was not able to pro-vide the system with sufficient power to operate during the ship'sentire voyage. We do not recommend the use of solar panels in thisapplication, and it is necessary to use the ship power supply. Thesystem did not create any interference from other ship systems.

More specifically, this solution presented can assist the BrazilianPSC, which needs to control ships that have been operating in morethan 100 port terminals installed in coastline and fluvial basins. TheBrazilian coastline has an extension of 7,408 km and several portswith different water quality. Hereby, it can be applied in othercountries located in several parts of the world, as an efficient BWtreatment system has not been developed.

Brazil is divided in eight fluvial basins composed of big rivers,where some of them can receive ships from cabotage that conductBW operations. For example, in the La Plata, Uruguay and ParaguayRivers and the Patos Lagoon system, the famous Limnoperna fortune(golden mussel) was introduced by BW from ships (Uliano-Silvaet al., 2013). Depending on ship cabotage routes, they can moveinto these fluvial basins to load and unload cargo, where the BWdischarge may occur. This can be examined considering the shiproutes monitored in this investigation.

Ships that ballast in fluvial ports will probably deballast in seaports in the course of coast littoral or in fluvial ports located on theAmazon and Par�a rivers, in the northeast region of Brazil. Therefore,the control of the BW quality inside of the tanks is a great indicatorfor inhibition of transfer of invasive species for other regions in thiscountry. In this investigation, we identified that this ship has beenexchanging BW as recommended by Brazilian regulation.

5. Conclusion

We concluded that this system is able to monitor whether shipsare operating according to IMO recommendations, performing theoperation of BWE at least 200 nautical miles from the nearest landin water, at least 200 m' depth.

This investigation identifies that this cabotage ship performedBWE operations near the Brazilian coastline and also at least 50nautical miles from the nearest land in waters of at least 200 mdepth. This has been occurring due to ship routes that have a loweroperation cost for the ship. Some BWEs occurred in the Amazonregion when the ship was operating in Par�a and Maranh~ao Statesinside of the Amazon river or near its mouth, which was indicatedin the past for the Brazilian Navy in NORMAM 20/2005.

N.N. Pereira et al. / Ocean & Coastal Management 131 (2016) 25e38 37

In January 2014, the Brazilian Navy changed the previousregulation and published the NORMAM 20/2014, which removedthe obligation of a second BWE for cabotage ships' entrance into theAmazon region, and ships do not need to change their route toattend to the request of at least 50 nautical miles. This study canthen alert the IMO and PSC about the risk of transferring invasivespecies, because ships can ballast in rivers and deballast into thesame macro fluvial region in short voyages, as had occurred in theAmazon region. Thus, marine and fresh water species can be easilyadapted to new regions, since they are in the same macro regionwhere the water characteristics are similar.

On the other hand, the BWmonitoring system presented is ableto reduce mistakes in the BWRF as well as to provide more reli-ability in the information given to Port State Control than thosereports presented by the ship crews in the paper form. Our resultsshow that ship correctly conducts the BWE, as informed in theBWRF. The graphic interface developed to read data sent by thesystem can be easily adapted to receive data from the acquisitionsystem and to convert them into a BWRF. Herewith, the PSC canreceive BWRF information with more reliability with inclusion ofBW quality parameters.

Furthermore, this system can generate cost reductions for shipballast inspection operations from PSC. Ship BW quality data can bereceived remotely before the ship arrives at port areas. In fact, theBW remote monitor inside of the ship tanks is a great challenge tobe implemented on a large scale. Several problems can affect theefficiency of this system, such as power energy supply using solarpanels, data transmission, sensor reading quality and data inter-pretation, ship tank access and the sensors' maintenance interval.

Evenwith these difficulties, we evaluated the BWE operation byphysical-chemical parameters variation and identified the correctposition for BWE. This allows us to conclude that salinity is usuallythe main parameter in determining the BWE, but depending on thewater quality, the changing in turbidity can also be a good indicatorof whether there has been monitoring in the BW. Sensors with highsensitivity can indicate small variations, which can be comparedwith BWRF information. Usually, the turbidity is not indicated witha BWE indicator, even in the BWM Convention. Thus, if ships have aBWRMS onboard, port authorities can use the turbidity to validatethe BWE.

Acknowledgement

We would like to acknowledge Norsul, a Brazilian shippingcompany, for giving us this great opportunity to install the BWmonitoring system in theM/V Crateus. The ship's crewwas essentialduring the installation process. This study was financed and sup-ported by the Brazilian National Council of Scientific and Techno-logical Development - CNPq (558151/2009-4) and the BrazilianInnovation Agency e FINEP (0112016800). We specially acknowl-edge Prof. Dr. Rui Carlos Botter for contacts with ship companies toinstall our system, as well as naval engineer Geert Jan Prange, whohelped us conceive this system, and Alexandre Tavares Lopes tohelp us in this development and ship installation.

References

Affonso, A.G., Queiroz, H.L.D., Novo, E.M.L.D.M., 2011. Limnological characterizationof floodplain lakes in mamirau�a sustainable development reserve, centralAmazon (Amazonas state, Brazil). Acta Limnol. Bras. 23 (1), 95e108.

Agency, NHS, 2003. Brazil-Ballast Water Anvisa e GGPAF Projects 2002. NationalHealth Surveillance Agency.

Alcantara, E., Novo, E., Stech, J., Lorenzzetti, J., Barbosa, C., Assireu, A., Souza, A.,2010. A contribution to understanding the turbidity behavior in an Amazonfloodplain. Hydrol. Earth Syst. Sci. 14 (2), 351e364.

Brown, C., 2012. Patterns of management, compliance, and geography of ballastwater in California. Mar. Invasive Species Program Calif. State Lands Comm.

2012 Prevention First Meeting. (p.27).Carlton, J.T., Geller, J.B., 1993. Ecological roulette: the global transport of nonin-

digenous marine organisms. Science 261 (5117), 78e82.Caron Junior, A., 2007. Assessing the Risk of Introduction of Exotic Species in the

Port of Itajaí and Around through Ballast Water. Master Dissertation. Universityof Vale do Itajaí in Environmental Science and Technology, BR (in Portuguese).

Casco, S.L., Carnevali, R.P., Poi, A.S., Neiff, J.J., 2014. The influence of water hyacinthfloating meadows on limnological characteristics in shallow subtropical waters.Am. J. Plant Sci. 5 (13), 1983.

Choi, K.H., Kimmerer, W., Smith, G., Ruiz, G.M., Lion, K., 2005. Post-exchangezooplankton in ballast water of ships entering the San Francisco Estuary.J. Plankton Res. 27 (7), 707e714.

Clark, R.B., 1986. Marine Pollution. Oxford University Press.Cohen, A., 1998. Ships’ Ballast Water and the Introduction of Exotic Organisms into

the San Francisco Estuary: Current Status of the Problem and Options forManagement. San Francisco Estuary Institute, Richmond, CA, p. 90.

Cohen, A.N., Foster, B., 2010. The regulation of biological pollution: preventingexotic species invasions from ballast water discharged into California coastalwaters. Golden Gate Univ. Law Rev. 30 (4), 3.

Cohen, A.N., Dobbs, F.C., 2015. Failure of the public health testing program forballast water treatment systems. Mar. Pollut. Bull. 91, 29e34.

Commission (California State Lands Commission), 2014. Assessment of the Efficacy,Availability, and Environmental Impacts of Ballast Water Treatment Technolo-gies for Use in California Waters, p. 2014.

Doblin, M.A., Murphy, K.R., Ruiz, G.M., 2010. Thresholds for tracing ships' ballastwater: an Australian case study. Mar. Ecol. Prog. Ser. Vol. 408, 19e32.

Dobroski, N., Scianni, C., Gehringer, D., Falkner, M., 2009. 2009 assessment of theefficacy, availability and environmental impacts of ballast water treatmentsystems for use in California waters. Prod. Calif. State Legis p.100.

Drake, L.A., Ruiz, G.M., Galil, B.S., Mullady, T.L., Friedmann, D.O., Dobbs, F.C., 2002.Microbial ecology of ballast water during a transoceanic voyage and the effectsof open-ocean exchange. Mar. Ecol. Prog. Ser. 233, 13e20.

Duggan, I.C., Van Overdijk, C.D., Bailey, S.A., Jenkins, P.T., Lim�en, H., MacIsaac, H.J.,2005. Invertebrates associated with residual ballast water and sediments ofcargo-carrying ships entering the Great Lakes. Can. J. Fish. Aquatic Sci. 62 (11),2463e2474.

Gollasch, S., 2006. A new ballast water sampling device for sampling organismsabove 50 micron. Aquat. Invasions 1 (1), 46e50.

Gray, D.K., Johengen, T.H., Reid, D.F., MacIsaac, H.J., 2007. Efficacy of open-oceanballast water exchange as a means of preventing invertebrate invasions be-tween freshwater ports. Limonol. Oceanogr. 52, 2386e2397.

Great Lakes Seaway, 2015. 2014 Summary of Great Lakes Seaway Ballast WaterWorking Group.

Hallegraeff, G.M., 1992. Harmful algal blooms in the Australian region. Mar. Pollut.Bull. 25 (5), 186e190.

IMO, 2004. International Convention for the Control and Management of Ships'Ballast Water and Sediments.

Jewett, E.B., Hines, A.H., Ruiz, G.M., 2005. Epifaunal disturbance by periodic lowlevels of dissolved oxygen: native vs. invasive species. Mar. Ecol. Prog. Ser. 304,31e44.

Johengen, T.H., Reid, D.F., MacIsaac, H.J., Dobbs, F.C., Doblin, M.A., Fahnenstiel, G.L.,Jenkins, P.T., 2005. Assessment of Transoceanic Nobob Vessels and Low-salinityBallast Water as Vectors for Nonindigenous Species Introductions to the GreatLakes. Final report for the project to the Great Lake Protection Fund. the Na-tional Oceanic and Atmospheric Administration, the US Environmental Pro-tection, and the US Coast Guard. Available from: www.glerl.noaa.gov/res/.

Johengen, T.H., MacIsaac, H., Dobbs, F., Doblin, M., 2007. Identifying, Verifying, andEstablishing Options for Best Management Practices for NOBOB Vessels.

Leal Neto, A.C., 2007. Identifying Similarities: an Application to Risk Assessment ofBallast Water. Thesis (Ph.D.) submitted to the Federal University of Rio deJaneiro. Department of Science in Energy Planning, BR (in Portuguese).

Morrison, G., Fatoki, O.S., Persson, L., Ekberg, A., 2001. Assessment of the impact ofpoint source pollution from the Keiskammahoek Sewage Treatment Plant onthe Keiskamma River-pH, electrical conductivity, oxygen-demanding substance(COD) and nutrients. Water Sa. 27 (4), 475e480.

Murphy, K.R., Field, M.P., Waite, T.D., Ruiz, G.M., 2008. Trace elements in ship'sballast water as tracers of mid-ocean exchange. Sci. Total Environ. 393, 11e26.

Murphy, K., Ruiz, G.M., Sytsma, M., 2003. Standardized Sampling Protocol forVerifying Mid-ocean Ballast Water Exchange. DTIC Document.

Pereira, N.N., Brinati, H.L., 2012. Onshore ballast water treatment: a viable option formajor ports. Mar. Pollut. Bull. 64 (11), 2296e2304.

Pereira, N.N., Botter, R.C., Folena, R.D., Pereira, J.P.F.N., da Cunha, A.C., 2014. Ballastwater: a threat to the Amazon Basin. Marine Pollut. Bull. 84 (1), 330e338.

Peterlin, M., Kontic, B., Kross, B.C., 2005. Public perception of environmental pres-sures within the Slovene coastal zone. Ocean Coast. Manag. 48 (2), 189e204.

Piola, A.R., Matano, R.P., Palma, E.D., M€oller, O.O., Campos, E.J., 2005. The influenceof the Plata River discharge on the western South Atlantic shelf. Geophys. Res.Lett. 32 (1).

Prange, G.J., Pereira, N.N., 2013. Ship ballast tank sediment reduction methods. Nav.Eng. J. 125e12, 127e134.

Quay, P.D., Wilbur, D., Richey, J.E., Devol, A.H., Benner, R., Forsberg, B.R., 1995. The18O: 16O of dissolved oxygen in rivers and lakes in the Amazon Basin: deter-mining the ratio of respiration to photosynthesis rates in freshwaters. Limnol.Oceanogr. 40 (4), 718e729.

Ruiz, G.M., Carlton, J.T., Grosholz, E.D., Hines, A.H., 1997. Global invasions of marine

N.N. Pereira et al. / Ocean & Coastal Management 131 (2016) 25e3838

and estuarine habitats by non-indigenous species: mechanisms, extent, andconsequences. Integr. Comp. Biol. 37, 621e632.

Smith, L.D., Wonham, M.J., McCann, L.D., Ruiz, G.M., Hines, A.H., Carlton, J.T., 1999.Invasion pressure to a ballast-flooded estuary and an assessment of inoculantsurvival. Environ. Res. 1 (1), 67e87.

Tamburri, M.N., Wasson, K., Matsuda, M., 2002. Ballast water deoxygenation canprevent aquatic introductions while reducing ship corrosion. Biol. Conserv. 103(3), 331e341.

Torgan, L.C., 2011. Flowering of algae: composition, causes and consequencesINSULA. J. Bot. 19, 15e33 (in Portuguese).

Uliano-Silva, M., Fernandes, F.C.F., de Holanda, I.B., Rebelo, M.F., 2013. Invasivespecies as a threat to biodiversity: the golden mussel Limnoperna fortuneapproaching the Amazon River basin. Explor. Themes Aquat. Toxicol. Kerala135e148.

Vandermeulen, J.H., 1996. Environmental trends of ports and harbours: implicationsfor planning and management. Marit. Policy Manag. 23 (1), 55e66.

Vazquez, H.P., Contento, L., Bachur, J., Ingallinella, A.M., Sanguinetti, G., 1998.Removal of micro-organisms by clarification and filtration processes: nationalreport Argentina. 21 st Int. water Serv. Congr. Exhib. 16 (1).