A spatial temporal decision framework for adaptation to sea level rise

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55

ENSO2912_proof ■ 30 March 2013 ■ 1/13

Environmental Modelling & Software xxx (2013) 1e13

Contents lists available at SciVerse ScienceDirect

Environmental Modelling & Software journal homepage: www.elsevier.com/locate/envsoft

A spatial temporal decision framework for adaptation to sea level riseq Q3

Oz Sahin a, b, *, Sherif Mohamed a a

Q1

b

Griffith School of Engineering, Griffith University, QLD, Australia South East Queensland Climate Adaptation Research Initiative, Australia

a r t i c l e i n f o

a b s t r a c t

Article history: Received 17 August 2012 Received in revised form 10 March 2013 Accepted 11 March 2013 Available online xxx

There is a strong link between decision making and environmental stresses. Two dilemmas confront decision makers: how and when to adapt to sea level rise, due to complexities of environmental systems and the changing nature of the decision making process. This process is inherently complex and often involves many stakeholders with conflicting views. Considering the complexity and dynamic nature of coastal systems, this paper introduces a Spatial Temporal Decision framework to assess coastal vulnerability, and the adaptation alternatives to SLR. The STD is based upon a combination of: System Dynamics modelling; Geographical Information Systems modelling; and multicriteria analyses of stakeholders’ views using the Analytical Hierarchy Process. For case study analyses, the City of the Gold Coast located in Southeast Queensland, Australia has been selected. The results of the vulnerability assessment indicate that, at the end of a 100 year simulation period, approximately 6% of the landscape in the study area will be gradually inundated over time, with 0.5 cm rise per year. However, the percentage of the vulnerable area leapt to about 34% for Scenario 2, and 56% for Scenario 3, which represent 1 cm and 1.5 cm rise per year. Using the information obtained from vulnerability assessments, three stakeholder groups (Politicians, Experts and Residents) were consulted to determine the goal, criteria and adaptation alternatives for the multicriteria analyses. Analyses of survey data reveal that across the three stakeholder groups, Effectiveness and Sustainability are the criteria of highest priority. Ó 2013 Elsevier Ltd. All rights reserved.

Keywords: Decision making Dynamic modelling Sea level rise Vulnerability Adaptation

1. Background There is overwhelming scientific consensus over the causes and impacts of climate change (IPCC, 2007). Sea level rise (SLR) is one of the most recognized possible impacts of changing climate. Coastal areas are economically productive and three times more densely populated than the global average (Small and Nicholls, 2003). Clearly, while communities have benefited from the many advantages of living in these areas, inevitably they also face the threat of natural disasters and specifically from SLR via permanent inundation of low-lying regions, inland extension of episodic flooding, increased beach erosion and saline intrusion of aquifers (McLean et al., 2001).

q This manuscript is the extended version of the conference paper presented at 2012 IEMSs conference in Leipzig. * Corresponding author. Griffith School of Engineering, Griffith University, QLD, Australia. E-mail addresses: o.sahin@griffith.edu.au, [email protected] (O. Sahin). URL: http://www.csiro.au/partnerships/seqcari

Coastal communities have been adapting to changing conditions throughout history. However, faced with increased threats due to SLR, coastal communities must act faster to develop more effective management policies. Moreover, the impacts of SLR are not expected to be spatially uniform across the world (Nicholls et al., 2007). It is, therefore, essential for decision makers (DM) to consider the dynamic and spatial characteristics of these changes in assessing the impacts of SLR when making decisions about the future. There is a range of analytical tools available to improve decision makers’ (DM) ability to understand and evaluate environmental management problems such as simulation models, GIS, and experts systems. However, although these tools provide invaluable information for decision making, each tool addresses only one aspect of a management problem. Therefore, effective decision making, in a dynamic complex environment, requires the expansion of the mental modelling boundaries and the development of additional tools to help DMs better understand how complex systems behave. Thus, DMs need to integrate each tools’ analytical results into a rational choice about what to do, where to do it, and when to do it (Schmoldt, 2001).

1364-8152/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.envsoft.2013.03.004

Please cite this article in press as: Sahin, O., Mohamed, S., A spatial temporal decision framework for adaptation to sea level rise, Environmental Modelling & Software (2013), http://dx.doi.org/10.1016/j.envsoft.2013.03.004

56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110

111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175

ENSO2912_proof ■ 30 March 2013 ■ 2/13

2

O. Sahin, S. Mohamed / Environmental Modelling & Software xxx (2013) 1e13

1.1. Addressing the vulnerability The Intergovernmental Panel on Climate Change (IPCC) defines vulnerability as: “the degree to which a system is susceptible to, and unable to cope with, adverse effects of climate change, including climate variability and extremes. Vulnerability is a function of the character, magnitude, and rate of climate change and variation to which a system is exposed, its sensitivity, and its adaptive capacity” (Parry et al., 2007). This definition incorporates three main variables: (1) exposure; (2) sensitivity; and (3) adaptive capacity. Thus, vulnerability is a function of the exposure, sensitivity and adaptive capacity of a system. Based on the IPCC definition, many researchers (Dongmei and Bin, 2009; Eriyagama et al., 2010; Rosenzweig and Tubiello, 2006; Webersik et al., 2010; Yusuf and Francisco, 2009) present vulnerability in the form of the following function: Fig. 1. Five dimensional STD framework.

V ¼ ðE; S; ACÞ

(1)

Where, V ¼ Vulnerability, E ¼ Exposure, S ¼ Sensitivity, and AC ¼ Adaptive Capacity. In Eq. (1), changes over time and space are not explicitly taken into account. Hence, vulnerability assessments, based on the above definition, have assumed that vulnerability is a static process and, so, have been conducted with reference to a target year and sea level rise prediction (DCCEE, 2009; Wang et al., 2010; Wu et al., 2009). However, vulnerability is a dynamic process and should be considered as a dynamic continuum. Further, as the three elements constituting vulnerability interact with each other and change in time and space, so does the vulnerability. To include the dynamic and spatial aspects of vulnerability, by considering the time and space dependency, vulnerability can be expressed as:

Vðt; sÞ ¼ fðEðt; sÞ; Sðt; sÞ; ACðt; sÞg

(2)

Where; t: Temporal dimension, s: Three spatial dimensions (x,y,z). V(t,s), in the above definition (Eq. (2)), represents four dimensions in describing the vulnerability of a particular system, region or group with respect to time and space; it changes as time passes. The ultimate goal of vulnerability assessment is to predict the vulnerability and provide information and guidance to DMs. Therefore, the use of an appropriate assessment approach would strengthen the DMs’ abilities to take appropriate action with speed and accuracy. Evidently, dealing with environmental problems requires an approach that can take into account human decision making processes, as well as time and space. Therefore, integrating vulnerability assessments with decision making process would constitute the fifth dimension of vulnerability and adaptation analyses and, as a result, enhance DMs ability to take concrete action towards adaptation to climate change. 2. Approach In the light of the above discussion, this paper introduces a Spatial Temporal Decision (STD) framework to assess coastal vulnerability, and the adaptation alternatives to SLR. The STD approach takes into account five dimensions of the decision process in coastal areas (Fig. 1). Space (x,y,z) and time (t) constitute the first four dimensions, and provide a common base where all natural and human processes occur. This approach is crucial in generating adequate information from which DMs can devise realistic adaptation strategies. For this reason, it is essential to incorporate the first four dimensions into the fifth dimension, the element of human decision making (h).

Thus, developing STD is based upon a combination of: System Dynamics (SD) modelling; Geographical Information Systems (GIS) modelling; and multicriteria analyses of stakeholders’ views using the Analytical Hierarchy Process (AHP) (Saaty, 1980). The cyclic STD process consists of: 1) Identification of the problem; 2) Vulnerability assessment by using Dynamic Spatial Model (DSM) which combines a spatial model (GIS) and a temporal model (SD); 3) Evaluating potential adaptation strategies by using a Multiple Criteria Decision Aid (MCDA) approach, based on information obtained from the previous step; 4) Test the solution using the dynamic simulation model to see if the preferred adaptation strategies are adequate enough to provide an effective solution; and 5) Refine the model to eliminate or reduce any weakness in the solution. 2.1. Understanding and treatment of uncertainties There is a range of uncertainty types, often resulting from incomplete scientific understanding of the various processes. Indeed, some uncertainties are caused by the processes themselves, which function in space and time, and so cannot be captured by the models. As a consequence, the uncertainties cannot be reduced. The adequate treatment of uncertainties is a crucial aspect in the development and application of integrated assessment models. According to Robinson (2003), uncertainty about the future stems from three sources: 1) The lack of knowledge about system conditions and underlying dynamics; 2) The prospects for innovation and surprise, and 3) The nature of human decision-making. From a positivist perspective, the first two sources of uncertainty can be overcome and reduced by progressively improving the models and including more variables (Mannermaa, 1991). However, system theorists question the validity of the positivist approach, arguing that it is theoretically impossible to pre-state all possible future outcomes for most real systems. According to them, behaviour of complex systems is non-linear, chaotic, and sensitively dependent on the initial conditions and contexts, which are not fully known (Kay et al., 1999). Swart et al. (2009) provide a useful review of how uncertainty has been treated in the assessments of the IPCC, and how this treatment has evolved over time. Firstly, uncertainty, as described in the IPCC Third Assessment Report, occurs as a result of problems with: The data (i.e. errors in the data, random sampling errors, and biases); The models (i.e. the processes are known, but the functional relationships or errors in the structure of the model are unknown; the structure is known, but the values of some important parameters are unknown or are erroneous; the historical data and

Please cite this article in press as: Sahin, O., Mohamed, S., A spatial temporal decision framework for adaptation to sea level rise, Environmental Modelling & Software (2013), http://dx.doi.org/10.1016/j.envsoft.2013.03.004

176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240

ENSO2912_proof ■ 30 March 2013 ■ 3/13

3

O. Sahin, S. Mohamed / Environmental Modelling & Software xxx (2013) 1e13

241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305

model structure are known, however, the reasons for believing in the parameters or the model structure will change over time); and Other sources of data or information (i.e. the use of inappropriate spatial/temporal units; uncertainty due to projections related to human behaviour; and the ambiguously defined concepts and terminology, etc.). Although uncertainty is an essential aspect in policy and decision making processes, a formal treatment of uncertainty is difficult to achieve. Hence, in recognising the potentially critical influence of uncertainties, the authors have adopted a variety of measures in designing the model in a two phase step. In the first phase, uncertainties surrounding the vulnerability assessment models have been addressed through the use of scenarios, sensitivity analysis, and performance tests. In the second phase, to address the temporal and spatial uncertainties surrounding the decision process, the model output of vulnerability analysis was incorporated into the MCDA analysis, then a sensitivity analysis is conducted to further minimise the uncertainties. 2.2. Model development Depending on the rate of the SLR, an area that is now subject to a 1 in 100 year flood risk may, in time, and with a high enough SLR, face more frequent flood events, or become permanently inundated (McInnes et al., 2000). As a result, the boundary of the coastal flood plain will shift inland over time. Thus, the ability to identify lowlying areas is a crucial factor in any coastal region vulnerability assessment. Further, coastal inundation and flooding, stemming from SLR and associated extreme events, can be modelled by establishing the interactions between time and space to capture the changes in a coastal system. Importantly, the physical processes such as overland flow, and proximity to and connectivity of the area with neighbouring areas should be considered essential for modelling inundation. The direction of the flooding, between the adjacent gridcells, depends on the difference in elevation between them. In order to capture the fundamental dynamic processes of inundation the area under consideration is subdivided into a cellular (i  j) grid to simulate how flood water spreads between adjacent cells. Each cell represents a specific area corresponding to one of four cover types: Sea, Waterways, Pond, or Land. The transition rules, which describe the relationships between the cells and the criteria showing how the states of a cell are to change, regulate the behaviour of the system. Based on the following equation, the flood water diffusion from one cell to another is projected:


 F Xi;j ¼



processing systems capable of storing, retrieving and displaying various type of data. Given the strength of SD in representing the temporal processes, especially with restricted spatial modelling capabilities, the association of SD and GIS produces a synergy effect. As a result, addition of a spatial dimension to SD would enable modellers to explicitly; 1) simulate system structure that is heterogeneous over space, as well as 2) consider how spatial interactions affect systems themselves (BenDor and Kaza, 2012). As spatial detail improves model accuracy, visualisation and usability for simulating changes in time and space, a number of researchers have proposed the use of a versatile approach, which considers many aspects of the problem by combining GIS with SD (Ahmad and Simonovic, 2004; Gharib, 2008; Grossmann and Eberhardt, 1992; Ruth and Pieper, 1994; Sahin and Mohamed, 2010; Zhang, 2008). In this paper, therefore, the authors employ a DSM approach by loosely coupling GIS and SD approaches. The DSM approach utilises GIS to reduce the data preparation and processing workloads considerably for use in the SD model, enhance the spatial visualisation, and reveal spatial relationships. Meanwhile, the SD model is employed to deal with the dynamics of the complex system, revealing its causal structure and the relations of the system component while adding temporal dimension to the spatial modelling capability of GIS. The DSM consists of three components: SD (temporal) model, GIS (spatial) model, and the data convertor (Fig. 2). The DSM captures the changes in time and space by obtaining and processing the temporal data from the SD and the spatial data from the GIS by exchanging data through the data convertor. There are three common approaches for coupling GIS and SD; loose coupling, moderate integration and tight integration (Gimblett, 2002; Maguire et al., 2005). There are trade-offs that are unique to each of the coupling strategies, since each approach has its advantages and disadvantages. Although the loose coupling approach has some disadvantages, such as slow execution speed and low simultaneous execution capability, this approach is adopted to link SD and GIS in this study by considering some of its overpowering advantages, such as (Fedra, 2006): ease of use, both in GIS and SD can be modified and run without any complication; data structures do not have to be matched; data can be transformed to each other’s formats through a converter; users are able to make on-the-fly changes more rapidly, and It is fast and portable; the SD model can be used with different GIS.





  1 CT Xi;j ¼ L and dcCT Xn;m ¼ W; and CE Xi;j  dcCE Xn;m 0; otherwise

Where, F is, either flooded (1) or not flooded (0); CE is the cell elevation; CT (Xi,j) is the cover type, either inundated L or not inundated W; CT (Xn,m) is the adjacent cells cover types, either L land (or other cover types other than sea) or W sea (or became sea due to inundation); X represents a grid cell; (n,m) refers to all adjacent cells to i,j (i.e.: i,j 1, i,jþ1, iþ1,j and i 1,j). 2.2.1. Dynamic Spatial Model (DSM): linking GIS with SD SD allows us to model, describe and better understand the behaviour of complex systems, and thus, to identify and manage the information related to these systems. GIS are spatial data

(3)

2.2.1.1. Temporal model component. The temporal modelling segment consists of the building, integration and running of two types of models: an Inundation Model and a Vulnerability Model. Fig. 3 illustrates the overall model structure and its submodels. These submodels of the temporal components interact with each other through feedback links. The Sea level and Population variables are the two key drivers affecting Inundation and Vulnerability Models. For the model, a hundred year time horizon is considered as from the 2010 through to 2110; this scale is consistent with most SLR scenarios developed by the IPCC (Meehl, 2007).

Please cite this article in press as: Sahin, O., Mohamed, S., A spatial temporal decision framework for adaptation to sea level rise, Environmental Modelling & Software (2013), http://dx.doi.org/10.1016/j.envsoft.2013.03.004

306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370

ENSO2912_proof ■ 30 March 2013 ■ 4/13

4

371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435

O. Sahin, S. Mohamed / Environmental Modelling & Software xxx (2013) 1e13

Fig. 2. Dynamic spatial model structure.

For building the temporal model, the Vensim DSS (Decision Support System) software was chosen because of its flexibility when representing continuous or discrete time, a graphical interface, or performing causal tracing, optimization, and sensitivity analysis (Ventana Systems, 2012). 2.2.1.1.1. Inundation Model. Fig. 4 shows the inundation model structure and some of the variables assumed to be important in a coastal system, and their interactions. As seen in Fig. 4, the system comprises three state variables; Cell Cover, Elevation and Sea Level. The Sea Level is an exogenous variable causing changes in both the Elevation and Cell Cover variables, over time. The model assumes a linear increase in the Sea Level over time, based on a range of SLR projections ranging from 0.5 m to 1.5 m. The SLR, at a given time, is calculated by:

SLt ¼

Z

dR*dt þ SL0

(4)

Where: SLt represents the linear Sea Level at time t, dR is the rate of rise at each time step dt, and SL0 denotes the initial Sea Level at the beginning of the simulation.

Fig. 3. Model structure and its submodels

For modelling purpose, the study area is subdivided into a cellular grid to simulate how flood water spreads between adjacent cells, based on the conceptual framework for inundation. This grid is then superimposed over the coastal area. Each cell represents a specific area corresponding to one of four cover types: Sea, Waterways, Pond, or Land. At each simulation step, the state of each cell is determined by the condition of its neighbours to the north, east, south, and west. At each simulation step, as the sea level rises, the elevation of a cell is determined by its condition at the previous time step, its border conditions with its neighbours, and the cover type of its neighbours (Land, Waterways, Sea, or Pond). The elevation of a cell is determined by adjusting the elevation, at the previous time steps, by the flow-in (Increase) and the flow-out (Decrease) of the cell, according to the properties of the adjacent cells. The Elevation is the integral of the net flow of Increase and Decrease, which is mathematically represented by the following equation:

Et ðx; yÞ ¼

Zt

ðIt ðx; yÞ

Dt ðx; yÞÞdt þ E0 ðx; yÞ

(5)

0

Where: Et (x,y) denotes cell elevation at location (x,y) at a given time; E0 (x,y) represents initial cell elevation at location (x,y); It (x,y) shows the rate of elevation increase at location (x,y); and Dt (x,y) indicates the rate of elevation decrease at location (x,y). The changes in cell elevation occur when only the Cover Type of a cell is Land, Waterways, or Pond, at time step tn, and it is transformed into Sea at the next time step, tnþ1. Here, the cell is assumed to be inundated from the rising sea level and, therefore, the elevation of the cell is updated, and said to be equal to the Sea Level at the time period tnþ1. The Cell Cover is the integral of the net flow of Change and Change Previous. Based on the following equation, the type of Cell Cover in any given time is determined.

Please cite this article in press as: Sahin, O., Mohamed, S., A spatial temporal decision framework for adaptation to sea level rise, Environmental Modelling & Software (2013), http://dx.doi.org/10.1016/j.envsoft.2013.03.004

436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500

ENSO2912_proof ■ 30 March 2013 ■ 5/13

5

O. Sahin, S. Mohamed / Environmental Modelling & Software xxx (2013) 1e13

501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565

Fig. 4. Inundation model based on cell elevation and cover types.

CTt ðx; yÞ ¼

Zt 0

ðCt ðx; yÞ


 CPt ðx; yÞÞ*dt þ CT0 x; y

(6)

Where: CTt (x,y) represents cell cover type at location (x,y) at a given time; CT0 (x,y) indicates initial cell cover type at location (x,y); Ct (x,y) denotes rate of cell cover type change at location (x,y); and CPt (x,y) denotes the rate of previous cell cover type change at location (x,y). As the model runs, the state of the each cell is assessed simultaneously. The Change flows into the cell (Stock) and updates the Cover Type of the cell for the present time step (i.e. t1). Subsequently, Change Previous removes the Cover Type of the cell at the previous time step (i.e. t0). This is necessary to assign only one Cover Type value to the cell for each time step. For example, if the Change alters the Cover Type of a cell from Land to Water at time step (t1), then Change Previous discards the previous cover type value (Land) from the cell. 2.2.1.1.2. Vulnerability Model. Vulnerability to SLR results from a combination of various factors, such as high population density along the coast, and the susceptibility of coastal regions to coastal storms, as well as other effects of climate change. Therefore, an accelerated SLR could fundamentally change the state of the coast

and, as a result, coastal environments and human populations will be affected significantly. In the final building step of the temporal model component, the Vulnerability model is developed to estimate the potential impacts of SLR (Fig. 5). The critical vulnerability of coastal areas to coastal storms (in the short term) and SLR (in the long term) relates to flooding. Therefore, the vulnerability assessment (VA) needs to focus on people and properties. Hence, two VA indicators are selected: 1) People at Risk over time due to coastal flooding, and 2) Area at Risk (loss of land) due to inundation and coastal flooding. First, the number of people who live in the area is calculated based on two stocks in the model: The Population (P0) that resides in the area at the beginning of simulation, and the Residents (Rt), which is the integral factor of the Population Increase (Pt). The model determines the changes in the population living in the area using the following equation:

Rt ðx; yÞ ¼

Zt 0


 Pt ðx; yÞdt þ P0 x; y

(7)

Where: Rt (x,y) indicates people reside at location (x,y) at a given time; P0 (x,y) denotes initial number of people reside at location

Fig. 5. Vulnerability model for people and property at risk.

Please cite this article in press as: Sahin, O., Mohamed, S., A spatial temporal decision framework for adaptation to sea level rise, Environmental Modelling & Software (2013), http://dx.doi.org/10.1016/j.envsoft.2013.03.004

566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630

ENSO2912_proof ■ 30 March 2013 ■ 6/13

6

631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695

O. Sahin, S. Mohamed / Environmental Modelling & Software xxx (2013) 1e13

(x,y); and Pt (x,y) represents rate of population increase at location (x,y). Then, People at Risk are calculated by multiplying the sum of Flooded Cells with the Cell Size:

VPt ¼

X

Rt ðx; yÞ*FCt ðx; yÞ*Cs

(8)

Where: VPt (x,y) is vulnerable people at a given time; Rt (x,y) denotes people residing at location (x,y) at a given time; Cs represents a constant value showing size of each grid cell; and FCt (x,y) shows flooded cells at location (x,y) Then, the Area at Risk is calculated by multiplying the sum of the Flooded Cells with the Cell Size:

At ¼

t X 0


 FCt x; y *Cs

(9)

Where: At (x,y) is vulnerable area at a given time; Cs represents a constant value showing size of each grid cell; and FCt (x,y) shows flooded cells at location (x,y) 2.2.1.2. Spatial model component. Spatial analysis is a set of methods whose results change when the locations of the objects being analysed change (Longley, 2005). Importantly, spatial analysis derives information from the data using the spatial context of the problem and the data. Spatial modelling involves the use of disaggregated spatial data and relationships in order to understand spatial forms and process (BenDor and Kaza, 2012). That is, it deals with space. GIS is the main tool used in the spatial analysis. In this study, the ArcInfo 9.3.1 is used to develop the spatial model (ESRI, 2009), which is connected to the simulation model through the data convertor and file monitor application developed for this framework by the authors. There are two main ways to spatially model sea level rise and subsequent coastal inundation. Geospatial data depicts the real world in two forms, which leads to two distinct approaches: the object-based model, and the field-based model (Goodchild, 1992). The object-based method uses contour lines; it is usually suitable for a very rapid and simple risk assessment over large areas. However, it does not take into account the presence of intervening topographic ridges or other features (e.g. man-made defences) that can separate a low-lying area from the source of flooding (Brown, 2006). That is, since a contour-line method relies solely on elevation data, inaccuracies arise when deriving a vulnerable zone based on this method because it does not consider connecting cells. The Raster model, as Lo and Yeung (2007) define it, is one of the variants of the field based models of geospatial data modelling. It is best employed to represent spatial phenomena that are continuous over a large area. For example, the Raster data model uses a regular grid to cover the space; the value in each cell represents the characteristic of a spatial phenomenon at the cell location. In computing algorithms, a raster can be treated as a matrix with columns and rows (xey coordinates), and its values can be stored into a 2D array. These characteristics hence make integration of GIS and SD easier, especially since SD can easily use array variables for data manipulation, aggregation, and analysis. Therefore, the raster data model was selected for spatial modelling. The basic elements of a raster model include the cell value, cell size, raster bands, and spatial reference (Chang, 2006). Each cell in a raster has a value (integer or floating) representing the characteristic of a spatial phenomenon at the location denoted by its column and row (x,y). Depending on the data type, both integer and floating point rasters are used in spatial modelling. For example, the research considers a sea level rise of 0.5e1.5 cm. Thus, a floating-point raster is more suitable for the

elevation data, as rise of sea level represents continuous numeric data with decimal digits, i.e. 10.125 m, 10.124, and so forth. However, the integer values are used for land cover rasters, i.e. 1 for Sea, 2 for Waterways, 3 for Pond, and 4 for Land. Essentially, the cell size determines the resolution of the raster model. As a larger raster cannot provide the precise location of the spatial features, the model result may not be satisfactory. Nevertheless, the smaller cell size can address these problems; although their use increases the data volume and data processing time, considerably. There are always trade-offs between the quality of the model outcomes and the processing time. In this study, a 5 m cell size is used for the modelling. The elevation data are the most critical elements in assessing the potential impacts of rising sea level. The uncertainty of the elevation data affects the delineation of the coastal elevation zones. Most elevation datasets have vertical accuracies of several metres or even tens of metres (at the 95% confidence level). Gesch (2009) argues that the mapping of submeter increments of sea-level rise is highly questionable, especially if the elevation data used have vertical accuracy of a metre or more (at the 95%). That is, the elevation uncertainty is much smaller for the more accurate elevation data. To keep the analysis reasonably manageable, this study has focused on the vertical accuracy. Therefore, to acquire more accurate results, the research used 5 m DEM with 0.1 m vertical accuracy. A variety of data from different sources is required as inputs to the spatial model. All the data layers needed to be in grid (raster) format, with a resolution of 5  5 m cell size. By working at a high spatial resolution, the model is able to reflect, accurately, the spatial changes in inundation resulting from the SLR. This approach provides a convenient way for describing the geo-processing procedure in GIS. Hence, based on this approach, we begin by converting the shape-files to the raster format, then reclassifying, and correcting their projection and, finally, unifying the coordinate system by using the model builder (Fig. 6). The vector data are, consequently, converted to a raster format. The Model Builder is a graphical tool for automating a model through the use of a work flow. Spatially, the size of the raster cell generated was based on the minimum mapping unit (5  5 m) to match the DEM data. The attribute assignments are based on the centroid of the cell. Australian Bureau of Statistic (ABS) data on dwellings and the Digital Cadastral Database (DCDB) are also converted to a raster format. Uncertainty, however, exists regarding where the population resides within the census parcel. Therefore, in the current study, the vulnerable population is estimated as a

Fig. 6. Vector to raster conversion using the model builder.

Please cite this article in press as: Sahin, O., Mohamed, S., A spatial temporal decision framework for adaptation to sea level rise, Environmental Modelling & Software (2013), http://dx.doi.org/10.1016/j.envsoft.2013.03.004

696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760

761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825

ENSO2912_proof ■ 30 March 2013 ■ 7/13

O. Sahin, S. Mohamed / Environmental Modelling & Software xxx (2013) 1e13

percentage of the census population, based on the inundated parcels. 2.2.2. Data convertor The loose coupling approach involves the transfer of data between the GIS and SD. Hence, it is necessary to establish, create and manipulate data files, so that they can be exported or imported between the spatial and temporal components of the model. The data in the files can be stored in several file formats. Different file formats have different characteristics, depending on a range of factors, such as the source of the data, and the software architecture. As the STM combines two different modelling approaches, it is useful to choose a device independent file format which can be usable by both applications, regardless of their hardware or software platforms. Therefore, in the current study, the device independent ASCII file format for GIS, and the .cin and tabtext file formats for SD are chosen for the cross-platform exchange of data. When exchanging data between two applications, it is necessary to convert the data formats into the right file format, as used by the applications (i.e., ASCII /.cin, and/or .cin / ASCII). To assist with this process, a converter program is developed. The converter program involves two separate applications: the data converter and the file monitor. The data converter software automates the format transition between the ArcGIS and SD data formats. First, it converts the ArcGIS text (ASCII) files to SD text files (.cin), and then it converts the files from the SD .tab files back to the ArcGIS .txt files. All code for the data converter is written in Cþþ under Visual Studio 2008, using the Microsoft.NET framework version 2.0. As a console application, it takes its commands via program arguments. 2.2.3. Decision model Decision making is a process of selecting from among several alternatives, based on various (usually conflicting) criteria. Information on priority alternatives is vital in aiding DMs to design more effective adaptation options and better management plans to reduce the adverse effects of SLR. The current study will use the MCDA technique because it is the most suitable approach by which to identify the priority of adaptation alternatives. Several multicriteria decision aid techniques are suitable for comparing multiple criteria, simultaneously, and for providing a solution to a given problem. While there are no better or worse techniques, some techniques are better suited to a particular decision problem (Haralambopoulos and Polatidis, 2003). The AHP technique, despite some criticisms, has been selected for the current study. Criticisms of the AHP include; Difficulty of conversion from verbal to numeric scale; Inconsistencies imposed

7

by the 1 to 9 scale; Number of comparisons required may be large (Macharis et al., 2004; Ramanathan and Ganesh, 1995). However, AHP, owing to its flexibility to be integrated with different techniques, enables the user to extract benefits from all the combined methods and, hence, achieve the desired goal in a better way (Vaidya and Kumar, 2006). In addition, the AHP is set apart from other MCDA techniques because of the unique utilisation of a hierarchy structure to represent a problem in the form of a goal, criteria and alternatives (Saaty and Kearns, 1985). This allows for a breakdown of the problem into various parts for pair wise comparisons, which uses a single judgement scale. Thus, the AHP has been widely used to solve various decision problems. Examples of the recent AHP applications include: Awasthi and Chauhan (2011); Bottero et al. (2011); Crossman et al. (2011); Chen and Paydar (2012); Do et al. (2012); and Gao and Hailu (2012). The underlying concept of the AHP technique is to convert subjective assessments of relative importance to a set of overall scores or weights (Saaty, 1980). When making decisions it is necessary to ensure that the alternatives selected for further consideration/implementation are consistent with the value systems and preferences of the stakeholders (Basson and Petrie, 2007). The large body of the literature emphasises the importance of stakeholder participation in the decision making process on environmental issues, which are complex, uncertain, and vary in time and space (Adger et al., 2007; Lim et al., 2004; van den Hove, 2000; Willows and Connell, 2003). Jakeman and Letcher (2003) also emphasise the importance of stakeholders’ involvement in the model validation process. Although there are many definitions, in the literature, for the term stakeholder, in the current study, the following IPCC definition is used as it refers to (Parry et al., 2007): ‘people or organisations, who have an investment, financial or otherwise, in the consequences of any decisions taken’. Here, three key stakeholder groups, within the study area, were included in the decision making process: Expert, Residents, and Politicians. To achieve and facilitate a workable process to reduce the vulnerability of an area and a population to SLR, a hierarchical (AHP) structure was designed (Fig. 7). The hierarchical elements, goal, criteria and alternatives are identified and finalised through stakeholders’ consultation in the study area. The specific goal used in the AHP structure is to reduce SLR vulnerability. To clarify further, this goal implies the identification and evaluation of adaptation alternatives in an attempt to reduce the negative impacts from SLR. It encompasses the idea behind the entire effort to reduce the negative impacts from climate change, specifically SLR. The criteria used in the decision making process to evaluate the alternatives, with respect to the goal, were: Applicability, Effectiveness, Sustainability, Flexibility, and Cost. The five chosen adaptation alternatives

Fig. 7. AHP structure.

Please cite this article in press as: Sahin, O., Mohamed, S., A spatial temporal decision framework for adaptation to sea level rise, Environmental Modelling & Software (2013), http://dx.doi.org/10.1016/j.envsoft.2013.03.004

826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890

891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955

ENSO2912_proof ■ 30 March 2013 ■ 8/13

8

O. Sahin, S. Mohamed / Environmental Modelling & Software xxx (2013) 1e13

were: Retreat, Improve Building Design, Improve Public Awareness, Build Protective Structures, and Take No Action. 3. Implementing the approach For case study analyses, the City of the Gold Coast located in Southeast Queensland, Australia has been selected. The area encompasses a diverse range of features including sandy beaches, estuaries, coastal lagoons and artificial waterways and is highly vulnerable to SLR. In this region, the maximum tidal range is 1.8 m, and on average, the coast is affected by 1.5 cyclones each year (Boak et al., 2001). Many of the residential areas in the city are filled to the 1:100 year flood level (Betts, 2002). 3.1. Vulnerability assessment To determine the effect of changes in vulnerable populations and land areas over time, the Cover Type and Elevation data were simulated under a number of SLR. The changes were captured in an SD model and exported to a GIS model for visualisation. The inundation

layer was overlayed with the 2001 Australian Bureau of Statistics census data, which was aggregated by census parcel for the area. Fig. 8 presents a series of flood maps generated by the model. It shows the extent of the areas at risk due to rising sea level, over a period of 100 years. Clearly, as inundation occurs at the watereland interface, the land area in close proximity to the sea, and around water bodies, were identified as the most vulnerable areas. The rising sea quickly penetrates inland through waterways and submerges the vulnerable areas around them, thus, putting the people currently living in those areas at risk. As shown in Fig. 9, at the end of a 100 year simulation period, approximately 6% of the landscape in the study area will be gradually inundated over time, with 0.5 cm SLR per year. Importantly, a 0.5 cm SLR does not pose any significant threats to the local population. However, this situation dramatically changes with scenarios 2 and 3, which represent 1 cm and 1.5 cm SLR per year. Indeed, the percentage of the vulnerable area leapt to about 34% for Scenario 2, and 56% for Scenario 3. The most noticeable changes occur after the first 25 years. Further, the rate of inundation becomes much higher after the first 50 years of the simulation period for both scenario 2 and scenario 3.

Fig. 8. Flood maps generated by the model.

Please cite this article in press as: Sahin, O., Mohamed, S., A spatial temporal decision framework for adaptation to sea level rise, Environmental Modelling & Software (2013), http://dx.doi.org/10.1016/j.envsoft.2013.03.004

956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020

1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085

ENSO2912_proof ■ 30 March 2013 ■ 9/13

O. Sahin, S. Mohamed / Environmental Modelling & Software xxx (2013) 1e13

Fig. 9. Area at Risk and Population at Risk under three SLR scenarios.

Although a substantial fraction of the landscape is threatened by the rising SLR, the percentage of the population that can be classified as vulnerable is relatively low for Scn2 and Scn3 scenarios, only 0.5% and 7%, respectively. The answer lies with most of the population residing at high altitudes. Nevertheless, the population located near waterways and coastal strips was especially vulnerable. Indeed, about 6% of the study area landscape will be submerged if the sea level rises a 0.5 m by 2110 (Fig. 9). Hence, the area at significant risk will be increased, up to 34% and 56% with a 1 m and 1.5 m rise in sea level, respectively. However, the inundation will, generally, be restricted to fringing shorelines and finger waterways margins (Fig. 8). Additionally, although, up to 56% of the land area will be facing the risk of inundation, the impacts of the same SLR scenarios on the residential areas are much smaller. 3.2. Multi-criteria decision analysis for adaptation options The fifth dimension of the current framework focuses on linking vulnerability assessment with the evaluation of adaptation alternatives through the use of AHP. The implementation of the MCDM models involved: assigning weights and priorities to the criteria by stakeholders; the normalisation of the raw scores to create a common scale of measurement; and the calculation of the decision

9

scores used to generate the final output from the models. By using the questionnaire, the participants were asked to compare the relative importance of the decision alternatives pairwise, with respect to criteria and the goal. The results were obtained through the use of Expert Choice11 package for computing relative weights, consistency ratio and local and global priorities (Expert Choice, 2008). Additionally, the MS Excel 2007 was also been employed for some calculations and data plotting. The AHP allows the inconsistency of every participant’s survey responses to be represented by the consistency ratio (CR). The resulting CRs are 0.02 for Residents, 0.02 for Experts and 0.06 for Politicians e all less than the 10% limit. The result indicates that stakeholder groups’ judgements with respect to each criterion and the goal are expected to be highly consistent. As seen in Fig. 10, regarding the Residents, from the five different adaptation alternatives presented in the survey questionnaire, the highest priority alternative was Improve Building Design (0.325 priority), closely followed by Build Protective Structures (0.285 priority). The least preferred alternative was Take No Action, followed by Retreat, with priorities of 0.061 and 0.102 respectively. In contrast, the Experts gave their highest priority to Improve Public Awareness with priority of 0.289, while Improve Building Design and Retreat were deemed the next most important alternatives with priorities of 0.278 and 0.203, respectively. While in accord with the Residents judgements for their least preferred alternative (Take No Action had a 0.089), the experts next least preferred alternative was Build Protective Structures (0.141 priority). The Politicians top two preferred adaptation alternatives were Improve Building Design with a priority of 0.457 (the Residents had this alternative as their top priority, while the Experts rated it as their second priority), and Retreat with a priority of 0.254, which was one of the Residents least preferred alternatives, but the Experts third top priority (Refer to the Politicians’ row in Fig. 10). Once again, the least preferred option for all three groups was Take No Action; however, the Politicians rated, as second to last, the alternative to Build Protective Structures, which disagreed with the Residents judgement, but agreed with the Experts judgement. The criteria priorities were obtained in the same way as the alternative priorities (Fig. 10). From the combined results for each stakeholder group, the two most important criteria to consider when making a judgement to reduce the negative impacts of SLR are Effectiveness and Sustainability. It appears that the three stakeholder groups uniformly agree about the importance of the criteria. For example, Applicability and Flexibility generally rank next highest (with Politicians the exception), while Cost ranks the lowest (with Politicians the exception ranking Flexibility last).

Fig. 10. Global criteria and alternative priorities for stakeholders.

Please cite this article in press as: Sahin, O., Mohamed, S., A spatial temporal decision framework for adaptation to sea level rise, Environmental Modelling & Software (2013), http://dx.doi.org/10.1016/j.envsoft.2013.03.004

1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150

1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215

ENSO2912_proof ■ 30 March 2013 ■ 10/13

10

O. Sahin, S. Mohamed / Environmental Modelling & Software xxx (2013) 1e13

3.3. Model refinement As there is a strong intersection between human decision making and environmental stresses, due to the uncertainty of the precise behaviour of complex environmental systems, two dilemmas confront DMs: how and when to adapt to SLR. While most decisions inherently are flawed to some degree, the decisions still have to be made. In order to improve the chance of better decision making, therefore, a robust process that considers the range of risks and associated uncertainties is required. The temporal model, introduced above, takes into account key variables that predict the extent and timing of coastal inundation. However, no variable was available to represent the adaptation alternatives. Thus, the model simulations were conducted under the Take No Action strategy. By modifying the model, 14 successive simulations were performed, with various values, to explore the impact of the Build Protective Structure and Improve Building Design adaptation options on vulnerable people and areas, as seen in Fig. 11. To be consistent with the modelling framework, the refined model tested the findings of the decision analysis. First, to test the efficiencies of Build Protective Structure, the model was modified by adding a variable to represent an imaginary protective structure along the shoreline. The term “Build Protective Structure” refers to coastal engineering activities that reduce the risk of flooding and inundation. The heights of the protective structure varied from 0 to 2.5 m to estimate the most effective height that provided the best protection. The imaginary wall was built by altering the initial elevations of the border cells whose initial cover types were Land and adjacent to cells with Sea. Secondly, to test the efficiency of the Improve Building Design option, the model was further modified by adding another variable (Improve Building Design). A comparison was made of the usefulness of building a 1 m or 2 m high protective structure to reduce vulnerable areas to a 1.5 cm SLR per year. The results showing vulnerabilities under three adaptation scenarios; No Action, 1 m and 2 m Protective Structure can be seen in Fig. 12. The findings show that building protective structures along the coastline does not have any effect on reducing the extent of the inundation under Scenario 3, and, therefore, does not reduce the vulnerability. Similarly, the overlapping lines also

0

1.5 Rise Rate

indicate that Protective Structures (both, 1 m and 2 m) will not provide any safeguard for the vulnerable population from rising sea level (Fig. 12). Secondly, to test the efficiency of the Improve Building Design option, the model was further modified by adding another variable. The newly added variable, the Improve Building Design option covers a wide range of adaptation measures, including (but not limited to) flood proofing, elevated building design, and minimum flood level. As it was not possible to test each adaptation measure under this category, the focus was specifically on two measures: elevated building design and minimum flood level. Further, it was assumed that new building regulations would be introduced, and that all existing and new buildings would be modified and/or designed accordingly. Based on these assumptions, the initial elevation of each cell with a Land cover type was increased by 1 m, and then 2 m. In contrast to Protective Structures, the Improved Building Design adaption option provides the vulnerable population with better protection (Fig. 13). As seen in Fig. 13, with a 1.5 m SLR over a 100 year period, 56% of the land area would be submerged. However, implementing the option Improved Building Design reduced the vulnerability down to 6.5%, and 0.1% for a 1 m and 2 m building elevation, respectively. It is clear that elevating structures by the amount of the SLR, or more, would keep these structures at the same elevation relative to the sea and, thereby, prevent their becoming more vulnerable as the sea level rises. Further simulations were conducted to compare the effectiveness of adaptation options by setting initial values for Rise Rate (max 1.5 cm/y) and Current 100 year SS Height (max 2.5 m). Using the simulation results, the impacts of the five adaptation options; No Action, Protective Structures (1 m and 2 m) and Improve Building Design (1 m and 2 m) were compared. The outcomes on vulnerable people and areas are shown in Table 1. Firstly, the Build Protective Structure adaptation option was not an effective strategy in reducing vulnerability to SLR and associated SS. Secondly, the presence of rivers and canals in the study area nullified the effectiveness of any protective structures against SS and SLR, especially when combined with heavy rainfall and flash flooding. Thirdly, as the sea level rises, flooding penetrates into the same places it has occurred before. However, the Improve

0 8 -4 4 0 3 Current 100 yr SS Height "Population Growth (%)" Protective Structure

0 2 Improve Building Design

Initial Sea Level

Rise Rate

Adapted Cell Improved Building Design

Sea Level Rise

Initial Elevation Increase

Protective Structure

Elevation Decrease

Initial Cover

Improve Building Design

Adapted Cell Protective Structure

Change

Cell Cover Change Previous

Adjacent Cells

Fig. 11. Modified temporal model with Improve Building Design variable.

Please cite this article in press as: Sahin, O., Mohamed, S., A spatial temporal decision framework for adaptation to sea level rise, Environmental Modelling & Software (2013), http://dx.doi.org/10.1016/j.envsoft.2013.03.004

1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280

1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345

ENSO2912_proof ■ 30 March 2013 ■ 11/13

11

O. Sahin, S. Mohamed / Environmental Modelling & Software xxx (2013) 1e13

Fig. 12. Vulnerable area and population to SLR, with and without Protective Structure.

Fig. 13. Vulnerable area and population to SLR with and without Improved Building Design.

Table 1 Comparing five adaptation alternatives under two scenarios. Simulation period

Adaptation alternatives Take no action

Protective structure e 1 m

Protective structure e 2 m

Improve building design e 1 m

Improve building design e 2 m

1.5 m SLR

1.5 m SLR þ storm surge

1.5 m SLR

1.5 m SLR þ storm surge

1.5 m SLR

1.5 m SLR þ storm surge

1.5 m SLR

1.5 m SLR þ storm surge

1% 3% 5% 9% 16% 26% 36% 43% 48% 56%

82% 88% 89% 90% 90% 91% 91% 92% 92% 92%

1% 3% 5% 9% 16% 26% 36% 43% 48% 55%

82% 88% 89% 90% 90% 91% 91% 92% 92% 92%

0% 0% 0% 0% 1% 1% 1% 2% 4% 7%

56% 75% 79% 81% 83% 85% 86% 88% 89% 89%

0% 0% 0% 0% 1% 1% 1% 2% 4% 7%

10% 19% 31% 38% 45% 51% 59% 72% 77% 80%

0% 0% 0% 0% 0% 0% 1% 1% 1% 7%

87% 89% 90% 91% 91% 92% 92% 92% 93% 93%

0% 0% 0% 0% 0% 0% 1% 1% 1% 7%

87% 89% 90% 91% 91% 92% 92% 92% 93% 93%

0% 0% 0% 0% 0% 0% 0% 0% 0% 0%

20% 41% 50% 59% 66% 76% 84% 88% 89% 90%

0% 0% 0% 0% 0% 0% 0% 0% 0% 0%

0% 0% 0% 1% 1% 2% 10% 29% 44% 55%

Scenarios 1.5 m SLR

1.5 m SLR þ storm surge

Area at risk (%) 2020 2% 82% 2030 4% 88% 2040 5% 89% 2050 9% 90% 2060 16% 90% 2070 27% 91% 2080 36% 91% 2090 43% 92% 2100 48% 92% 2110 56% 92% Population at risk (%) 2020 0% 87% 2030 0% 89% 2040 0% 90% 2050 0% 91% 2060 0% 91% 2070 0% 92% 2080 1% 92% 2090 1% 92% 2100 1% 93% 2110 7% 93%

Please cite this article in press as: Sahin, O., Mohamed, S., A spatial temporal decision framework for adaptation to sea level rise, Environmental Modelling & Software (2013), http://dx.doi.org/10.1016/j.envsoft.2013.03.004

1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 Q 21384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410

1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475

ENSO2912_proof ■ 30 March 2013 ■ 12/13

12

O. Sahin, S. Mohamed / Environmental Modelling & Software xxx (2013) 1e13

Building Design option offers a much better option against SS with a 1.5 cm/year SLR. As demonstrated above, this option has the potential to reduce, significantly, the vulnerabilities to a 1.5 m SLR. On the other hand, its shielding power diminishes against a 1.5 m SLR combined with SS. Thus, it can be concluded that the findings of the simulation are consistent with the AHP findings, that is, the Improve Building Design was ranked as the most preferred option by the Residents and Politician, while the Experts voted it the second most preferred option. However, the Politicians and Experts voted the Build Protective Structures as the fourth most preferred option, while the Residents ranked it as the second most preferred option. 4. Conclusion An innovative characteristic of the STD approach is its ability to evaluate the decision choices prior to their implementation. This is achieved by incorporating the DSM simulation results into the decision making process and, then, retesting the information, obtained from this process, using the DSM. The model’s ability to preevaluate decision choices is an important feature; its legacy is that communities can avoid or minimise their decision error, and increase their chance of obtaining better decisions. The STD approach, in summary, provides a critical tool for obtaining quantitative information for managing and making choices with the aim of effective decisions. This integrated approach has the capability to: (1) Generate important spatialtemporal information required by decision makers (DMs); (2) Provide new insights into complex coastal systems; (3) Address multicriteria decision problems involving multiple stakeholders; (4) Enable DMs to examine decision alternatives through the use of the Dynamic Spatial Model; and (5) Address uncertainties and generate alternative scenarios, based on different user inputs. Acknowledgements The authors gratefully acknowledge the funding from the Griffith Climate Change Response Program (GCCRP) and the Centre for Infrastructure Engineering and Management (CIEM) e Griffith School of Engineering. References Adger, W.N., et al., 2007. Assessment of Adaptation Practices, Options, Constraints and Capacity. In: Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of WG-II to the 4AR Assessment Report of the IPCC. Cambridge, UK. Ahmad, S., Simonovic, S.P., 2004. Spatial system dynamics: new approach for simulation of water resources systems. Journal of Computing in Civil Engineering, ASCE 18 (4), 331e340. Awasthi, A., Chauhan, S.S., 2011. Using AHP and DempstereShafer theory for evaluating sustainable transport solutions. Environmental Modelling & Software 26 (6), 787e796. Basson, L., Petrie, J.G., 2007. An integrated approach for the consideration of uncertainty in decision making supported by life cycle assessment. Environmental Modelling & Software 22 (2), 167e176. BenDor, T.K., Kaza, N., 2012. A theory of spatial system archetypes. System Dynamics Review 28 (2), 109e130. Betts, H., 2002. Flood damage analysis using GIS at Gold Coast Council. Australian Journal of Emergency Management 17 (1), 33e37. Boak, L., McGrath, J., Jackson, L.A., 2001. A case study: the Northern Gold Coast Beach Protection Strategy. In: Edge, B.L. (Ed.), 27th International Conference on Coastal Engineering, 2001 Sydney, Australia. American Society of Civil Engineers, pp. 3710e3717. Bottero, M., Comino, E., Riggio, V., 2011. Application of the analytic hierarchy process and the analytic network process for the assessment of different wastewater treatment systems. Environmental Modelling & Software 26 (10), 1211e1224. Brown, I., 2006. Modelling future landscape change on coastal floodplains using a rule-based GIS. Environmental Modelling & Software 21 (10), 1479e1490.

Chang, K.-T., 2006. Introduction to Geographic Information Systems. McGraw-Hill Higher Education, Boston. Chen, Y., Paydar, Z., 2012. Evaluation of potential irrigation expansion using a spatial fuzzy multi-criteria decision framework. Environmental Modelling & Software 38 (0), 147e157. Crossman, N.D., Bryan, B.A., King, D., 2011. Contribution of site assessment toward prioritising investment in natural capital. Environmental Modelling & Software 26 (1), 30e37. DCCEE, 2009. Climate Change Risks to Australia’s Coast: a First Pass National Assessment. Australia: The Department of Climate Change and Energy Efficiency. Do, H.T., et al., 2012. Design of sampling locations for mountainous river monitoring. Environmental Modelling & Software 27e28 (0), 62e70. Dongmei, J., Bin, L., 2009. Countermeasures of adaptation to climate change: establishment and application for implementation matrix. Ecological Economy 5, 102e111. Eriyagama, N., et al., 2010. Impacts of Climate Change on Water Resources and Agriculture in Sri Lanka: a Review and Preliminary Vulnerability Mapping. IWMI Research Report 135. Colombo, Sri Lanka: International Water Management Institute. ESRI, 2009. ArcInfo. ESRI, Redlands, CA. Arlington, VAExpert Choice, 2008. Fedra, K., 2006. Beyond GIS: Integrating Dynamic Simulation Models and GIS for Natural Resources and Environmental Management. Conference Proceedings of Middle East 2006. Dubai, UAE. Gao, L., Hailu, A., 2012. Ranking management strategies with complex outcomes: an AHP-fuzzy evaluation of recreational fishing using an integrated agent-based model of a coral reef ecosystem. Environmental Modelling & Software 31 (0), 3e18. Gesch, D.B., 2009. Analysis of lidar elevation data for improved identification and delineation of lands vulnerable to sea-level rise. Journal of Coastal Research, 49e58. Gharib, S., 2008. Synthesizing System Dynamics and Geographic Information Systems in a New Method to Model and Simulate Environmental Systems. PhD. University of Bergen. Gimblett, H.R., 2002. Integrating Geographic Information Systems and Agent-based Modeling Techniques for Simulating Social and Ecological Processes. Oxford University Press, Oxford ; New York. Goodchild, M.F., 1992. Geographical data modeling. Computers & Geosciences 18 (4), 401e408. Grossmann, W.D., Eberhardt, S., 1992. Geographical information systems and dynamic modelling: potentials of a new approach. Annals of Regional Science 26, 53e66. Haralambopoulos, D.A., Polatidis, H., 2003. Renewable energy projects: structuring a multi-criteria group decision-making framework. Renewable Energy 28 (6), 961e973. IPCC, 2007. Summary for policymakers. In: Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M., Miller, H.L. (Eds.), Climate Change 2007: the Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, UK and New York, NY, USA. Jakeman, A.J., Letcher, R.A., 2003. Integrated assessment and modelling: features, principles and examples for catchment management. Environmental Modelling & Software 18 (6), 491e501. Kay, J.J., et al., 1999. An ecosystem approach for sustainability: addressing the challenge of complexity. Futures 31, 721e742. Lim, B., et al., 2004. Adaptation Policy Frameworks for Climate Change. Cambridge University Press, Cambridge. Lo, C.P., Yeung, A.K.W., 2007. Concepts and Techniques of Geographic Information Systems. Pearson Prentice Hall, Upper Saddle River, NJ. Longley, P., 2005. Geographical Information Systems and Science, second ed. Wiley, Chichester. Macharis, C., et al., 2004. PROMETHEE and AHP: the design of operational synergies in multicriteria analysis. Strengthening PROMETHEE with ideas of AHP. European Journal of Operational Research 153 (2), 307e317. Maguire, D., Batty, M., Goodchild, M., 2005. GIS, Spatial Analysis, and Modeling. ESRI Press, Redlands, CA, USA. Mannermaa, M., 1991. In search of an evolutionary paradigm for futures research. Futures 23 (4), 349e372. McInnes, K.L., Walsh, K.J.E., Pittock, A.B., 2000. Impact of Sea Level Rise and Storm Surges on Coastal Resorts. CSIRO Tourism Research, Final Report. CSIRO Atmospheric Research, Aspendale, Victoria, Australia. McLean, R.F., et al., 2001. Coastal Zones and Marine Ecosystems. Climate Change 2001: Impacts, Adaptation, and Vulnerability Cambridge, UK and NY, USA. Meehl, G.A., 2007. IPCC Climate Change 2007: the Physical Science Basis: Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Nature Publishing Group. Nicholls, R.J., et al., 2007. Coastal Systems and Low-lying Areas. In: Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, UK. Parry, M., Canziani, O., Palutikof, J., 2007. Technical summary. Climate change 2007: impacts, adaptation and vulnerability. In: Parry, M.L., C., O.F., Palutikof, J.P., van der Linden, P.J., Hanson, C.E. (Eds.), Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental

Please cite this article in press as: Sahin, O., Mohamed, S., A spatial temporal decision framework for adaptation to sea level rise, Environmental Modelling & Software (2013), http://dx.doi.org/10.1016/j.envsoft.2013.03.004

1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540

1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559

ENSO2912_proof ■ 30 March 2013 ■ 13/13

O. Sahin, S. Mohamed / Environmental Modelling & Software xxx (2013) 1e13 Panel on Climate Change. Intergovernmental Panel on Climate Change 2007, Cambridge, UK. Ramanathan, R., Ganesh, L.S., 1995. Energy resource allocation incorporating qualitative and quantitative criteria: an integrated model using goal programming and AHP. Socio-economic Planning Sciences 29 (3), 197e218. Robinson, J., 2003. Future subjunctive: backcasting as social learning. Futures 35 (8), 839e856. Rosenzweig, C., Tubiello, F.N., 2006. Developing Climate Change Impacts and Adaptation Metrics for Agriculture. Global Forum on Sustainable Development on the Economic Benefits of Climate Change Policies, Paris, France. Ruth, M., Pieper, F., 1994. Modeling spatial dynamics of sea-level rise in a coastal area. System Dynamics Review 10 (4), 375e389. Saaty, T.L., 1980. The Analytic Hierarchy Process. McGraw-Hill, New York, NY. Saaty, T.L., Kearns, K.P., 1985. Analytical planning: the organization of systems Oxford. New York. Pergamon Press. Sahin, O., Mohamed, S., 2010. Coastal vulnerability to sea level rise: a spatiotemporal decision making tool. In: Industrial Engineering and Engineering Management (IEEM), 2010 IEEE International Conference on, 7e10 Dec. 2010, pp. 89e93. Schmoldt, D.L., 2001. The Analytic Hierarchy Process in Natural Resource and Environmental Decision Making. Kluwer Academic Publishers, Dordrecht ; Boston. Small, C., Nicholls, R.J., 2003. A global analysis of human settlement in coastal zones. Journal of Coastal Research 19 (3), 584e599.

13

Swart, R., et al., 2009. Agreeing to disagree: uncertainty management in assessing climate change, impacts and responses by the IPCC. Climatic Change 92 (1), 1e29. Vaidya, O.S., Kumar, S., 2006. Analytic hierarchy process: an overview of applications. European Journal of Operational Research 169, 1e29. van den Hove, S., 2000. Participatory approaches to environmental policy-making: the European Commission Climate Policy Process as a case study. Ecological Economics 33 (3), 457e472. Ventana Systems, 2012. In: Vensim DSS. 6.0b. Harvard, MA: Ventana Systems, Inc. Wang, X., et al., 2010. Coastal Inundation Under Climate Change: a Case Study in South East Queensland. Climate Adaptation Flagship Working Paper No. 6. CSIRO. Webersik, C., Esteban, M., Shibayama, T., 2010. The economic impact of future increase in tropical cyclones in Japan. Natural Hazards 55 (2), 233e250. Willows, R.I., Connell, R.K., 2003. Climate Adaptation: Risk, Uncertainty and Decision-making. UKCIP Technical Report. UKCIP, Oxford, UK. Wu, S.Y., Najjar, R., Siewert, J., 2009. Potential impacts of sea-level rise on the Mid and Upper-Atlantic Region of the United States. Climatic Change 95, 12e138. Yusuf, A.A., Francisco, H., 2009. Climate Change Vulnerability Mapping for Southeast Asia. Economy and Environment Program for Southeast Asia (EEPSEA), Singapore. Zhang, B., 2008. A Study of GIS-SD Based Temporal-spatial Modelling of Water Quality in Water Pollution Accidents. ISPRS Congress Beijing 2008, Proceedings of Commission II. Beijing Reed Business e Geo.

Please cite this article in press as: Sahin, O., Mohamed, S., A spatial temporal decision framework for adaptation to sea level rise, Environmental Modelling & Software (2013), http://dx.doi.org/10.1016/j.envsoft.2013.03.004

1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578