The total transport of the

The total transport of the mid-shelf flow evolves as follows. Progressing along the Central Channel pathway, the volume flux remains roughly constant as it GKPIPNPLLGLDST goes from CC to CSW approaching Hanna Shoal (order 0.5 Sv).3 East of here, where the flow bifurcates, there is roughly an equal split in transport around the two sides of the shoal. In particular, 42% of the transport progresses anti-cyclonically around the northern side of the shoal, and 58% of the transport progresses cyclonically around the southern side. Encouragingly, the sum of these two branches (eastward transport at HSN+southward transport at CSW=0.53 Sv) is nearly identical to that approaching the shoal (eastward transport at CSW=0.51 Sv). As these two branches flow around Hanna Shoal they split again into smaller filaments. On the southern side the split is nearly even in transport (0.16 Sv at CN versus 0.19 Sv at HSS) and roughly conserves mass (total of 0.35 Sv compared to 0.31 Sv before the split). On the northern side, the transport remains constant from HSN (0.22 Sv) to CSE (0.23 Sv), but then decreases in value at HSSE (0.11 Sv). This decrease likely arises because some of the flow is diverted to the outer-shelf or into the shelfbreak jet due to the presence of a small canyon upstream of HSSE. We note that there is a small amount of water (including winter water) flowing southward on the eastern end of this section (not shown), which has also been observed in larger amounts in other hydrographic surveys (R. Pickart, unpublished data). Beyond section HSSE the anti-cyclonic branch splits into two filaments, with most of the transport going into the northern filament.
Mass is conserved over the locus of the flow branches in Fig. 9 (within the uncertainty of the transport measurements), and we can account for the volume flux observed through the head of Barrow Canyon. At section BCH, 0.99 Sv of Pacific water is flowing northward through the canyon, which is approximately equal to the sum of the ACC transport in the coastal route and the transport in the Central Channel branch (together totaling 0.96, using the average value for CC, CN, and CSW). Furthermore, we can make inferences about the partitioning/origin of the flow through the canyon. As seen in Fig. 9b, 0.20 Sv is flowing northward on the western flank of the canyon at BCH, and 0.79 Sv is flowing northward on the eastern flank. This implies that the water flowing anti-cyclonically around the northern side of Hanna shoal (order 0.2 Sv) feeds the western flank of Barrow Canyon, while the combination of the water flowing cyclonically around the southern side of Hanna Shoal plus the ACC (which together sum to order 0.8 Sv) feeds the eastern flank. We note that this is consistent with the mean vectors from the Barrow Canyon mooring array which show southeastward flow emanating from south of Hanna Shoal to the eastern side of the canyon (Fig. 8).
The transport of winter water on the mid-shelf is partitioned similarly to the full transport of the branches (Fig. 9b), although mass does not need to be (nor is it) conserved over the whole domain because of the seasonal presence of this water mass. As was the case for the full transport, roughly half of the winter water from the Central Channel branch flows on either side of Hanna Shoal (roughly 0.25 Sv approaches the shoal, and 0.13 Sv continues east while 0.13 Sv is diverted south). The difference now is that, in the region where the filaments form, the transport of winter water is quite small. Upstream of here 46% of the total transport (on average) is due to winter water, which decreases to only 17% in the filaments. This suggests, as noted above, that in the 2011 survey this region was at the leading edge of the winter water, which is consistent with the small amount of water colder than ?1.6 °C observed flowing north in Barrow Canyon (most of the winter water in the coastal branch had likely passed through the canyon prior to the survey). Later in the season the winter water from the interior pathways around Hanna Shoal should begin draining through the canyon. This is in line with the observation that winter water is prevalent at section BCC during August but not July (Nobre et al., 2014).

Here we consider a model describing the cycling of

Here we consider a model describing the cycling of particles and thorium in two layers whose interfaces coincide with the depths of the SV PDK1 inhibitor2 traps deployed in 2005 at the DYFAMED site. Thus, the upper layer extends from 313 to 524 m, and the lower layer extends from 524 to 1918 m. We assume that in each layer the sinking flux of particles or of particulate thorium, fi(z)fi(z), varies linearly with depth, zz (counted positive downwards):equation(1)fi(z)=(zbot?z)fi(top)+(z?ztop)fi(bot)zbot?ztopwhere zbotandztop are depths of the bottom and top of the layer, respectively. We divide each flux, fi(z)fi(z), by its corresponding geometrically averaged settling velocities (Armstrong et al., 2009) to estimate the concentration of particles, of particulate 230Th, or of particulate 234Thequation(2)Ci(z)=(zbot?z)Ctop+(z?ztop)Cbotzbot?ztop
The model used in this study (Fig. 2 and Fig. 3) is almost identical to that used by Murnane, 1994, Murnane et al., 1994, Murnane et al., 1996 and Marchal and Lam, 2012, except that we separate particles into sinking velocity classes. It relies on samples collected using IRS SV sediment traps and has the advantage that particle sinking velocities were measured directly by the SV traps, avoiding the need to arbitrarily assign prior values for sinking rates for “large” particles. Additionally, since particles were separated by sinking speed instead of particle size, we have a direct measure of sinking velocity for “slow-sinking” particles, and do not need to rely on the assumption that “small” particles are “slow”, and that “slow” means “non-sinking”.
Fig. 2. “Two-layer” model describing thorium cycling. Dissolved thorium interacts with particulate thorium via adsorption, desorption, and remineralization of slow-sinking particles. In addition, dissolved thorium has a source from uranium radioactive decay and is lost via radioactive decay. Thorium on different sinking classes is exchanged via particle aggregation and disaggregation. Particulate thorium is lost through radioactive decay, vertical divergence, and desorption, and is gained by adsorption.Figure optionsDownload full-size imageDownload as PowerPoint slide
Fig. 3. “Two-layer” model describing particle cycling. Slow-sinking particles can form fast-sinking particles by aggregation; and fast-sinking particles can disaggregate to form slow-sinking particles. Slow-sinking particles are also lost by remineralization, which is ignored for fast-sinking particles.Figure optionsDownload full-size imageDownload as PowerPoint slide
In applying the two-layer model of Fig. 2, we incorporate thorium exchange between the dissolved phase and fast- and slow-sinking via adsorption and desorption, but ignore remineralization of fast-sinking particles. Specifically, dissolved 230Th and 234Th are produced from the radioactive decay of uranium and removed by radioactive decay. Dissolved thorium is also lost by adsorption onto sinking particles, and is added to solution by desorption from particles and remineralization from the slow-sinking particles. For these particles, 230Th and 234Th are lost by radioactive decay, desorption, and particle remineralization, aggregation, and sinking (when the sinking flux divergence is negative), and are added by adsorption and fast-sinking particle disaggregation. For the fast-sinking class, thorium is lost by radioactive decay, desorption, and disaggregation, and added by adsorption and the aggregation of slow-sinking particles. All interactions are assumed to have first-order reaction kinetics. All rate constants are assumed to be constant in a given layer, but are allowed to vary between the two layers.

We cannot evaluate our results separately

We cannot evaluate our results separately from a phylogenetic context. The ecological patterns in our study were dominated by significant increases in densities of elpidiid holothurians when food availability was very high, following from very low or undetectable densities of these species during a period of moderate supply. While two non-elpidiid species showed significant increases in density, other TH287 did not similarly exhibit unanimously higher densities in 2011-2014 compared to 2006-2009. Traits conferring ecological advantages to elpidiids during periods of high food supply, and conversely those that might limit maximum densities of other species, are likely to be phylogenetically constrained (Eckelbarger and Watling, 1995). The ability of elpidiids to select for (Iken et al., 2001) and monopolize consumption of fresh phytodetritus (Wigham et al., 2003a and Wigham et al., 2003b) through increased feeding activity during high-food flux events (e.g. Kaufmann and Smith (1997) and Bett et al. (2001)) leads to clear ecological advantages (Billett and Hansen, 1982, Billett, 1991 and Billett et al., 2010). Likewise their characteristic opportunistic reproductive strategies (Galley et al., 2008) involving small egg size (Hansen, 1975, Billett and Hansen, 1982, Tyler and Billett, 1988 and Wigham et al., 2003b), potentially short-duration larval stages (Wigham et al., 2003b), small body size (Gebruk et al., 2003), and potentially small size at recruitment (this study) might allow for accelerated generation times. Some holothurians outside of Elpidiidae might be physiologically limited in their ability to produce large numbers of eggs, as has been shown in O. mutabilis ( Ramirez-Llodra et al., 2005), which broods its young, and psychropotids, which produce exceptionally large eggs ( Hansen, 1975 and Tyler et al., 1985b). Rapid locomotion and feeding behavior have been predicted to confer trophic advantages to O. mutabilis exploiting detritus on the sea floor ( Roberts et al., 2003). These advantages did not translate to higher population densities of this species complex when fresh phytodetritus was available in great quantities at Sta. M. Any competitive advantage of elpidiids conferred by the characteristics above might be short-lived, or strongly dependent on the availability of fresh surface-derived nutrients. Elpidiids were found in very low or undetectable numbers when fresh phytodetritus was limiting, and they exhibited significant boom and bust cycles when it was abundant. From an evolutionary standpoint, if Elpidiidae forms a monophyletic group then many traits considered amenable to high holothurian density on the abyssal sea floor during boom times might have evolved once, in the most recent common ancestor of elpidiids. These traits would therefore represent a sample size of one in comparative studies (Garland et al., 1992). A detailed understanding of deep-sea holothurian phylogenetic relationships is required for studies using independent contrasts (Garland et al., 1992) to yield a more robust understanding of deep-sea holothurian responses to food supply.

Necrosulfonamide Starting from July GLAD LSS drifters were deployed around

Starting from July 20, 2012, 20 GLAD LSS drifters were deployed around the DWH location. It is interesting to see the LCSs and their related uncertainties near these drifters. In order to look at more detailed structures of LCSs in the areas of our interests and experiments at the same time, a much smaller domain is needed. Fig. 13 shows the same as Fig. 12, but with a smaller domain. Now, much more detailed LCS structures and uncertainties can be seen clearly from all four forecasts, particularly ncom1km, which displays well organized small scale repelling LCSs. Again, the detail structures of LCSs identified using numerical model output depend on the model resolutions.
Fig. 13. As in Fig. 12, but for the smaller domain around the GLAD drifters area.Figure optionsDownload full-size imageDownload as PowerPoint slide
Most of the drifters were more likely deployed in the troughs of LCSs; very few were on the ridges of LCSs. Since the repelling LCS acts as a transport barrier, if the drifter is on the ridge of the LCS, it Necrosulfonamide could fall to either side of the LCS. Once the drifter is on one side of the LCS, it will be trapped on that side, as it is almost impossible for a drifter to cross the LCS barriers. Therefore, the movement of a drifter depends greatly on how repelling LCSs change with time, and the drifter movement is generally constrained by the repelling LCSs. Accurately identified repelling LCSs will provide helpful guidance to drifter deployment, and to forecasting the trajectories of drifters (Lekien et al., 2005, Lermusiaux et al., 2006a, Coulliette et al., 2007 and Shadden et al., 2009).
As we discussed, this is only the initial step in applying the ensembles to repelling LCS studies. Our next step will be Necrosulfonamide to compare the predicted drifter trajectories, the identified LCSs and their associated uncertainties against the observed drifter trajectories from the CARTHE GLAD data set. It is expected that the full advantages of the ensembles will be demonstrated with this valuable source of drifter data from the CARTHE GLAD. We will report those results when they are available.
4. Discussion and conclusions
As the designated modeling team within CARTHE to support and provide guidance to the GLAD at-sea experiment in the summer of 2012, we have run several real-time ocean model forecasts, starting on May 16, 2012, well before the GLAD drifter deployment. These include two ensembles (gom32r and gom32q), ncom3km, ncom1km and hycom4km. All of these forecast outputs are archived and made available on web servers for all the CARTHE scientists and students involved in this project. The raw forecast data are also available if requested. The real-time forecast results were evaluated every cycle by the local scientists at NRL, and important information and findings were provided to CARTHE scientists in other organizations via emails or regular tele-conferences. The implementation and operation of these forecast systems were conducted successfully without a glitch and provided great value and real-time guidance to the drifter deployment. In this paper, we describe the details of these numerical forecast systems and the corresponding products including the RELO ensembles, particularly the calibrated ensemble. The introduction and description of these forecast systems will provide background to scientists inside and outside the CARTHE project.

Fig Time series of mean EKE anomalies and Yucatan transport

Fig. 9. Time series of mean EKE anomalies and Yucatan transport anomalies in (A) ITD1-3 and (B) BClim integrations. The range of the y-axis in panel (B) is half than in panel (A).Figure optionsDownload full-size imageDownload as PowerPoint slide
The LC transit is occasionally accompanied by the formation and detachment of large anticyclonic eddies, called Loop eddies or Rings, sized between 150 and 300 km in diameter. Their surface temperature signature is often lost in the seasonal heating of the upper water column, but sea surface anomalies are easily tracked by satellite altimetry (Leben and Born, 1993). Once formed, eddies move westward across the basin at a speed of approximately 2-5 km d?1 (Elliott, 1982 and Vukovich and Crissman, 1986), and persist for months to years until they AM 1172 decay through interactions with the continental shelf. The eddy shedding process and driving mechanism are not fully understood, and stochastic processes are likely to play a significant role (Nowlin et al., 2000 and Zavala-Hidalgo et al., 2006). Several hypotheses have been explored on the controlling factors of the LC shedding, focusing primarily on the role of the wind forcing (Chang and Oey, 2010b and Oey et al., 2003), of the Yucatan Channel transport strength (Bunge et al., 2002 and Ezer et al., 2003), of the variability in the circulation in the Florida Straits (Sturges et al., 2009), of the potential vorticity fluxes into the Gulf (Candela et al., 2002, Oey and Lee, 2002 and Oey, 2004). More recently, Lugo-Fernández (2007) concluded that the LC and its eddy-shedding behave as a nonlinear oscillator with a very short memory, with periodicity and amplitude linked to the North Atlantic Oscillation (NAO).
A detailed Loop ring separation analysis has been performed by Vukovich (2012) using sea surface temperature, ocean color, sea surface height, and in-situ data from ships in the Gulf of Mexico. Vukovich considered three periods: 1972-2010, 1972-2000, and 2001-2010 due to changes in the eddy shedding periodicity. The average separation period between two consecutive Rings is ten months over 1972-2010, nine months over 1972-2000 and eleven months for the period 2001-2010 (Leben, 2005, Sturges and Leben, 2000 and Vukovich, 2007). In particular, Vukovich (2012) identifies 13 shedding events between 2000 and 2008. Using the absolute dynamic topography and absolute geostrophic velocities maps (MADT) from Aviso, we find that 12 LC eddies separated from the main current. In the ITD and BClim runs we count 14, 13, 15, and 15 shedding episodes, respectively, and we conclude that a reasonable, even if slightly larger than observed, number of eddies detach from the LC in our simulations. Another characteristic of the LC shedding is that Loop Eddies can form in any month of the year (Vukovich, 2012) as shown in Fig. 10, where we compare ROMS integrations and the Aviso data.
Fig. 10. Distribution of the Loop Eddies shedded in the period 2000-2008 as function of their month of detachment.Figure optionsDownload full-size imageDownload as PowerPoint slide

Not only dominant phylum level lineages

Not only dominant phylum-level lineages, but also specific pelagic alpha- and gammaproteobacterial lineages, reappeared in post-spill TW-37 libraries: the SAR11 subclusters (Field et al., 1997; Fig. S5); the Arctic96BD-19 group of sulfur-oxidizing heterotrophs (Marshall and Morris, 2013) that is prevalent in stratified, oxygen-depleted conditions (Walsh et al., 2009); the uncultured AGG47 cluster associated with marine snow (DeLong et al., 1993); the uncultured North Sea ZD0417 cluster (Stevens and Ulloa, 2008), and the uncultured SAR156 lineage (Mullins et al., 1995) (Fig. 3). The widely distributed SUP05 lineage, a presumable sulfur oxidizer typical of oxygen-depleted water columns (Walsh et al., 2009 and Canfield et al., 2010), was found during and after the plume stage.
Fig. 3. Phylogeny of Gammaproteobacteria (Uncultured lineages, Cycloclasticus and methaneotrophs/methylotrophs) in the Gulf of Mexico water column near the Macondo wellhead, based on near-full length 16S rRNA genes. Clones from the pre-spill water column sample (March 10, 2010) are labeled “Prespill”; clones from surface oil slicks (May 5, 2010) are labeled “Surfaceoil”; clones from plume water column samples (May 31, 2010) are labeled “Plumeprofile”. Clones from September 12 and October 18, 2010, and from July 3, 2011, are labeled Postplume I, II and III, respectively. The clone designations are followed by sampling depth in meters, and a 3-digit clone ID ( Table S1). The scale bar corresponds to 10% sequence distance.Figure optionsDownload full-size imageDownload as PowerPoint slide
3.2. Pyrosequencing results for surface oil slick and the plume-impacted water column
The pyrosequencing results for the weathered oil mixture at the surface from May 5, 2010, and the water column samples of May 31, 2010 were broadly consistent with the 16S rRNA gene clone libraries for the same samples (Fig. 1), but in addition revealed bacterial populations that had remained undetected in the clone libraries (Table S3). In the surface sample, pyrosequencing representation for Cycloclasticus (>93%), Alteromonas (1.45%) and Pseudoalteromonas (1.2%) resembled the clone library results, whereas Colwellia and Halomonas were detected in smaller proportions ( Table S1). In contrast, the alkane-degrading DWH Oceanospirillales accounted for near 90 and 70% of the pyrosequencing reads in the two deep plume samples of late May 2010 ( Table S3).
The DWH Oceanospirillales pyrosequencing reads were congruent with full-length 16S rRNA gene clones of DWH Oceanospirillales from the Gulf of Mexico ( Redmond and Valentine, 2012) and from the Atlantic Ocean offshore North Carolina (D?Ambrosio, 2011), and formed at least three distinct phylogenetic clusters (Fig. 4). The pyrosequencing survey also validated a diverse community of hydrocarbon-degrading bacteria in the plume profile that went largely undetected in the clone libraries (Table S3). The PAH-degrading genus Cycloclasticus remained variably detectable throughout the water column. Psychrophilic heterotrophs of the genus Colwellia (the only group detected in the plume clone libraries besides the DWH Oceanospririllales) accounted for approx. 1 to 3% of the pyrosequencing reads within the plume. The alkane-degrading genera Oleiphilus and Oleispira were found in low abundances below and within the plume. The pyrosequencing representation of the TW-37 uncultured gammaproteobacterial groups (AGG47, Arctic96BD19, SUP05, ZD0417, SAR156) above and below the plume was strongly reduced within the plume ( Table S3). A similar trend was observed for Alphaproteobacteria. While SAR11 bacteria accounted for a tenth of the pyrosequencing fragments above and below the plume, their representation decreased within the plume ( Fig. 1). In general, pyrosequencing analysis indicated a functionally and phylogenetically diversified alpha- and gammaproteobacterial community in the hydrocarbon plume; pre-spill populations of uncultured bacteria and oil-degrading bacteria remained detectable against the dominant plume populations of DWH Oceanospirillales. This result is compatible with a complex functional gene repertoire of plume microbial communities sampled at the same time ( Lu et al., 2012).

Additionally the experiments reveal that oil carbon

Additionally, the experiments reveal that oil carbon is incorporated into Acalisib formed by coagulation of diatoms. Increased phytoplankton concentration (Hu et al., 2011) due to the accident itself, or due to mediating measures like the opening of the floodgates of the Mississippi and diversionary channels, contributed to the formation of exceptionally high concentrations of phytoplankton that potentially aggregated. Possible subsequent sedimentation events would be higher than usual and could transport large amounts of fossil carbon to depths. Moreover, experimental evidence suggests that the presence of oil increases the stability and cohesion of aggregates, implying faster and more efficient transport to depths.
Both types of marine snow, microbial snow or phytoplankton aggregates, as well as the formation of OMAs, presumably contributed towards transporting oil carbon to the seafloor after the DwH accident. Whereas high sedimentation rates of minerals would be associated with sinking OMAs, the organic matter content would be elevated when sinking marine snow dominates downward flux.
Understanding of the different conditions which led to the formation and sedimentation of oil residues will enable model projections of the spatial and temporal extent of sedimentation pulses after the DwH accident. Modeling efforts that combine data on the spatial distribution of oil, dispersants, minerals and phytoplankton with the mechanistic understanding of the formation mechanisms of oil-containing, sinking particles could hindcast the spatial and temporal extent of sedimentation events after the spill. These estimates can then be compared to measured sedimentation and accumulation rates.
The overall effects of Corexit 9500A on sedimentation of oil-rich marine snow is currently difficult to assess, in part because of the patchy Corexit delivery. Self-aggregation of Corexit may have occurred near the source, but mainly Corexit concentrations in the GoM were too low for self-aggregation to have been significant. However, even very low concentrations of Corexit inhibited marine snow formation in experiments. Likely, the presence of Corexit reduced microbial marine snow formation and sedimentation after the spill. The impact of Corexit on the formation of phytoplankton-oil aggregates remains to be determined. Obviously the inhibitory effect on the formation of sinking particles needs to be considered when evaluating the advantages and disadvantages of mediation by Corexit additions in the future.
The formation of sinking, oil-associated, particles has far reaching consequences for the fate and distribution of oil in marine environments. Marine snow are hotspots of microbial activity (Arnosti et al., 2014, Azam, 1998, Azam and Malfatti, 2007, McGenity et al., 2012 and Ziervogel et al., 2012), but also the vehicle for the efficient transport of enclosed material to depths.
The qualitative observation of high in situ concentrations of marine snow in May 2010, and the observation of high sedimentation rates in August-September 2010 (Passow, unpublished) suggest that the DwH spill may have directly and indirectly triggered several large sedimentation events. Such sedimentation events would transport hydrocarbons from the surface oil layer or the subsurface plume downward. During transit sticky marine snow scavenges other particles such as organisms, feces, and detritus with great efficiency ( Smetacek, 1985), potentially resulting in a marine snow “blizzard”, which could explain the extraordinarily high accumulation rates of material observed at the sediment surface. The mucus-rich character of microbial marine snow is consistent with the material found covering corals ( Hsing et al., 2013 and White et al., 2012) and the patchy distribution of this material on corals is consistent with sinking marine snow. Future spill response and assessment will have to take the possible formation of rapidly settling, oil-associated marine snow and its deposition on the seafloor into account.

The research reported in this paper extends our prior work

The research reported in this paper extends our prior work towards an efficient and effective mining framework. As illustrated in Fig. 1, the framework is divided into an event log pre-processing phase, a phase for integrated resource mining including cross-perspective patterns, and a model post-processing phase. We evaluate our approach with an implementation of the three phases; with simulation experiments for measuring performance; and with the application of the approach on a real-life event log for checking its effectiveness.
Fig. 1. Framework for discovering resource-aware, declarative process models.Figure optionsDownload full-size imageDownload as PowerPoint slide
This research extends our previous work [16] as follows: (i) the developed pre-processing method increases the efficiency of the approach; (ii) the developed post-processing techniques increase the understandability of the results; (iii) a prototype of the entire framework has been implemented using Drools; and (iv) the approach has been extensively validated. In addition, the mining approach is explained in more detail. With our work, we complement research on process mining with an extensive support of the organisational perspective.
The remainder of this paper is structured as follows: Section 2 introduces background information. Section 3 describes our process mining approach. 4 and 5 describe the event log preprocessing and postprocessing phases of the framework, respectively. Section 6 explains the evaluations performed. Section 7 describes the related work and Section 8 concludes the paper.
2. Background
In the following we introduce the concepts upon which our approach has been developed.
2.1. Organisational and cross-perspective patterns in processes
The well-known workflow resource patterns [14] capture the various ways in which resources are represented and utilised in business processes. Of specific interest to our research are the creation patterns since they OG-L002 describe different ways in which resources can be assigned to activities. These patterns, which will be referred to as organisational patterns from now on, include: Direct Distribution, or the ability to specify at design time the identity of the resource that will execute a task. Role-Based Distribution, or the ability to specify at design time that a task can only be executed by resources that have a given role. Organisational Distribution, or the ability to offer or allocate activity instances to resources based on their organisational position and their organisational relationship with other resources. Separation of Duties, or the ability to specify that two tasks must be allocated to different resources in a given process instance. Case Handling, or the ability to allocate all the activity instances within a given process instance to the same resource. Retain Familiar (a.k.a. Binding of Duties), or the ability to allocate an activity instance within a given process instance to the same resource that performed a preceding activity instance. Capability-Based Distribution, or the ability to offer or allocate instances of an activity to resources based on their specific capabilities. Deferred Distribution, or the ability to defer the specification of the identity of the resource that will execute a task until run time. History-Based Distribution, or the ability to offer or allocate activity instances to resources based on their execution history. Note that the creation patterns Authorisation and Automatic Execution are not in the list because lobe-finned are not directly related to resource assignment.
It has been identified that process control-flow is intertwined with dependencies upon resource characteristics [15]. For instance, sometimes an activity must be executed eventually before another one for specific resources but not OG-L002 for others. As an example, resources with a certain role (e.g., trainees) must always perform a certain activity (e.g., double-check result) before they can continue with the following activity, but this might not be required for other roles (e.g., supervisors). We call this pattern Role-Based Sequence.

Process ontology based approach to

3. Process ontology based approach to business process modeling
The process ontology based approach (POBA) captures semantics in a domain process ontology, intending to close the semantic gap between the modeling requirements and perception of process modelers. This is motivated by the fact that Cy3.5 hydrazide developing formal conceptual models from natural languages with semantic ambiguities may lead to severe logical errors.
The two major tasks of process modeling are: understanding domain knowledge; and developing models based upon domain knowledge. Semantic ambiguity can occur at two stages of model construction [38]: 1) from domain concepts to concepts in a modeler\’s mind; and 2) from a modeler\’s mind to process models. Concept ambiguity occurs when modelers attempt to understand domain concepts based upon user requirements. Construct ambiguity occurs when modelers develop constructs based upon their own understanding. The two examples introduced in the motivating example can be either concept or construct ambiguities. This research proposes to reduce semantic ambiguity in the manner shown in Fig. 1.
Fig. 1. POBA to easing semantic ambiguity in business process modeling.Figure optionsDownload full-size imageDownload as PowerPoint slide
POBA is divided into three phrases as shown in Fig. 2. Phase 1 addresses concept ambiguity; Phases 2 and 3 address construct ambiguity. In Phase 1, domain process ontologies are employed to formally represent domain knowledge, which can help modelers better understand the business domain and reduce concept ambiguity. To resolve construct ambiguity in Phase 2, domain concepts found in a domain process ontology are mapped to business process models in a systematic procedure. In Phase 3, an ontology-based validation approach identifies semantic problems, such as completeness and clarity. POBA is an iterative approach so modelers can go back to earlier phases to improve domain process ontologies or modeling constructs.
Fig. 2. Three phases of process ontology based approach.Figure optionsDownload full-size imageDownload as PowerPoint slide
3.1. Phase 1: Development of domain process ontology
The domain process ontology is a formal definition of domain knowledge amino acid sequence captures the essential components of business process models and is defined as a set of terms and relationships. Definition 1 defines basic concepts that are usually required by business process models.Definition 1.
Domain Process Ontology is defined as a tuple (T, R), where T is a set of terms and R is a set of relationships. Formally, Domain Process Ontology ::= (T, R).
Terms are classified as: role nouns, non-role nouns, and activity verb phrases [22] and [32]. Role nouns refer to the actors of specific business activities. Non-role nouns include names of products, services, and data, Activity verb phrases describe the activities to be performed by business roles. Fig. 3 formally defines terms, with TID the unique identifier. Terms consist of strings attached to a classification identifier such as “RT”, “NRT” or “AT” (“role term”, “non-role term” and “activity term”, respectively). For example, “seller” is represented as: (1, “seller”,“RT”).
Fig. 3. Definition of terms.Figure optionsDownload full-size imageDownload as PowerPoint slide
Relationships common across different domains (such as is_a) are basic relationships; others are domain-specific [23] and [47].
A basic relationship, defined in Fig. 4, where RID is the unique identification number. A typical basic relationship between terms “cash” and “payment” might be: (5, < cash >, “is_a”, < payment >). Based upon prior work in business process modeling [5], [14] and [32], we identify three types of advanced domain relationships:?Activity-performing relationship: connects two roles involved in an activity performed by one of the roles.?Temporal relationship: sequences activities performed by one role, e.g., “prior to”, “at the same time” and “mutually exclusive.”?Conditional relationship: captures conditions needed to perform one specific activity for a role; e.g., a bidder needs to pay, conditional on a successful bid.

The different concepts defined here are

The different concepts defined here are those depicted in Fig. 5.Dynamic manufacturing network (DMN): DMN is a network of partners implied in the collaborative development of a manufactured product, with associated applications supporting PLM process, system engineering processes and controlled urbanization of information system, with the solutions realizing the applications.DMN execution platform: It is a cross-organizational collaborative enterprise platform that acts as a DMN hub between the partners working on a given product, their extended enterprise processes, the application supporting these processes and the technologies that realize the applications. As a PLM hub, this ITF2357 hub is used for secured transportation of product and process data and provides different services related to standard-based exchange, sharing and linking of product and process data, and associated supporting systems (for design, production, operation or support).DMN modeling platform: It is a visual enterprise modeling platform relying on an enterprise visual modeling language. Such an enterprise modeling platform must allow the capture and presentation of interrelated motivation, business, applications and technologies, through the usage of views structured according as set of predefined viewpoints. The viewpoints are associated to DMN participants and their concerns for supporting system engineering, PLM and controlled urbanization of information system a consistent way. In particular, PLM processes, viewpoints associated to data architecture and controlled urbanization processes must be consistent.DMN workflow system: It ITF2357 is an enterprise workflow system, which must be extended in order to capture expected characteristics of participants of a cross-organizational workflow model, capturing the characteristics of actual DMN members which are plugged on the hub, comparing them and reporting when expected and actual characteristics do not match.DMN platform designer: It is a software design platform, which must be able to consume models coming from the DMN modeling platform, and extend them in order to provide PIM and PSM models, which will be used for generation of artifacts that will be deployed on the DMN execution platform.DMN development platform: It is a development platform allowing developing and deploying the different artifacts of a DMN execution platform from the DMN blueprints.DMN engineering platform: It is the applicative infrastructure for creation of an interoperable DMN environment, combining the DMN platform designer, the DMN development platform, the DMN modeling platform, the DMN workflow system, and the DMN execution platform, structured as a DMN software factory.DMN information structure viewpoint: It is a viewpoint comparable to the traditional information models created in the development of almost any information system. It shows the structure of the information used in the DMN organizations, in the PLM/Urbanization/SE business process and in supporting application and technologies. Stakeholders are manufacturing product and process data architects. The concern is the consistency, completeness and accuracy of data models and of domain-specific languages used in the DMN and by the DSF. It includes DSL and models provided by manufacturing, enterprise, business, applicative and ICT standards constituting the underlying federated interoperability framework [1] of the DMN.DMN software factory: It is an organization of a DMN Engineering platform aiming at taking advantage of model-driven architecture and model-driven engineering and realized through the usage of qualified model transformation capabilities based on open standards.DMN blueprint: It is a model of one or several nodes of a DMN. It is formalized in ArchiMate as a data object (application layer), realizing a business object which can be described using various representations, being active, passive or behavioral (as defined in ArchiMate 2.1). It is consequently a hypermodel [7]. Many realizations of a DMN blueprint exist at different places of the DMN infrastructure (ICT layer of the DMN). The list of nodes for which a DMN blueprint could be defined includes manufactured products, private business processes, business services, cross-organizational business processes, organizations, projects, personal, process segment, configuration items, methods, applications, software systems, devices, networks, plants, etc. Concerned business processes are PLM and SE processes. Concerned cross-organizational collaborative processes are primarily change and configuration management processes.