CANTERBURY REGIONAL COUNCIL v DEWHIRST LAND CO LTD AND ANOR [2019] NZCA 486 [8 October 2019]

CANTERBURY REGIONAL COUNCIL v DEWHIRST LAND CO LTD AND ANOR [2019] NZCA 486[8 October 2019]IN THE COURT OF APPEAL OF NEW ZEALANDI TE KŌTI PĪRA O AOTEAROACA34/2019[2019] NZCA 486BETWEEN CANTERBURY REGIONAL COUNCILAppellantAND DEWHIRST LAND COMPANY LIMITEDAND MICHAEL GRAHAM DEWHIRSTRespondentsHearing: 20 August 2019Court: French, Collins and Stevens JJCounsel: R J B Fowler QC and J L S Shaw for AppellantR J Somerville QC and J M van der Wal for RespondentsJudgment: 8 October 2019 at 11.30 amJUDGMENT OF THE COURTA The application for leave to bring a second appeal is granted.B The questions of law are answered as follows:First question:Did the High Court err in its assessment of the correct test for determiningthe extent of the riverbed in applying the definition of "bed" in s 2 ofthe Resource Management Act 1991?Answer:No.Second question:Did the High Court err in adding the phrase "usual or non-flood" into thedefinition of "bed" in s 2 of the Resource Management Act 1991 byimplication?Answer:Yes.Third question:Did the High Court err in concluding that the assessment of various flowrates or return periods was an irrelevant consideration in determining theextent of the riverbed?Answer:No.C The appeal is dismissed.D There is no order as to costs.____________________________________________________________________REASONS OF THE COURT(Given by Stevens J)Table of ContentsPara NoIntroduction [1]Factual Background [3]The application for leave to appeal [12]The High Court judgment [15]The first question of lawSubmissions of the Council [20]Respondents' submissions [32]Applicable statutory provisions [35]Our analysis [39]The common law [42]Summary of common law principles pertaining to rivers and beds [51]Development of New Zealand law on rivers, banks and beds [52]Bed of a river — statutory context [60]The second question of law [100]The third question of law [105]Result [108]Introduction[1] The respondents, Dewhirst Land Co Ltd (Dewhirst Ltd) and Mr Dewhirst werecharged with, and entered guilty pleas to, five charges against the ResourceManagement Act 1991 (the RMA) in the District Court.1 The charges were broughtby the appellant — the Canterbury Regional Council (the Council) — under sectionsof the RMA restricting certain uses of the riverbed and limiting the diversion of waterin a river.[2] This appeal is primarily concerned with the proper interpretation of "bed", asit relates to a river, under s 2 of the RMA. Relevantly, "bed" is defined as "the spaceof land which the waters of the river cover at its fullest flow without overtopping itsbanks". The Council contends that the High Court's interpretation (given on appealfrom the District Court) was incorrect.Factual background[3] The Selwyn River, known as a braided river, is located on the Canterburyplains. Dewhirst Ltd owns farming land adjacent to, and to the south of, the true rightbank of the river. Mr Dewhirst is a dairy farmer and a director of the company whoundertakes most of the day-to-day management. Dewhirst Ltd initiated a plan todevelop an area of land adjacent to the river, some of which is within the bed ofthe river.[4] Around May 2016, the respondents consulted the Council regarding whatconsents might be required to develop the land. Council officers indicated that someof the proposed development was within the flood control vegetation lines containedin the Council's Flood Protection and Drainage Bylaw 2013. It therefore met thedefinition of the "bed" of the river in the Canterbury Land and Water Regional Plan,as a result of which the proposed development works could not occur within the areaproposed. Such a view was not accepted by the respondents.1 Sections 13(1)(b) (two charges), 13(1)(a) (one charge), 13(2)–(2A) (one charge) and 14(2)(a) (onecharge) of the Resource Management Act 1991.[5] However, the respondents applied for, and were granted, a gravel authorisationwhich permitted the extraction of gravel from two sites in the Selwyn River with acombined total volume of 3,000 m³. Gravel was extracted from the river incircumstances which the Council considers exceeded the authorisation granted.[6] Later in 2016, vegetation within the same general area was cleared, up to thepoint where there was an existing formed bank. This was considered bythe respondents to be the edge of the riverbed. They also created a gravel bund alongthe line of that bank.[7] Upon pleading guilty to the charges in the District Court, the respondentsaccepted they had gone beyond the terms of the applicable authorisation and that theyhad, so far as the charges under s 13 of the RMA were concerned, encroached beyondthe line of the existing formed bank of the river. However, several aspects of thesummary of facts were not accepted. In particular, the respondents contested whetherthe entire bund and the area of vegetation comprising gorse and broom cleared waswithin the bed of the river. As Judge Hassan noted, the disputed matters relate to theextent of the offending, environmental harm and culpability (which are relevantfactors in applying sentencing principles). The proper interpretation of the definitionof "bed" in the RMA is central to this aspect of the dispute.2[8] The disputes regarding the summary of facts gave rise to two decisions inthe District Court. In the first decision, Judge Hassan resolved some of the disputedfacts, but the question of whether some or all of the development was within the bedof the Selwyn River was not determined.3 A second judgment followed whichfavoured the Council's interpretation that the definition of "bed" in the RMA meansthe space of land which the waters of the river cover with two conditions:4(a) first, at its fullest flow; and(b) secondly, without overtopping its banks.2 Canterbury Regional Council v Dewhirst [2018] NZDC 7650 at [3]–[4].3 Canterbury Regional Council v Dewhirst [2018] NZDC 5412.4 Canterbury Regional Council v Dewhirst, above n 2, at [61].[9] In so ruling the District Court Judge preferred evidence before him given byMr McCracken, a Chartered Professional Engineer river expert employed bythe Council as the Regional Lead for river engineering, to that of Mr Macky, the expertriver engineer for the respondents.5[10] The second judgment was appealed to the High Court where Gendall Jconsidered that Judge Hassan applied the wrong legal test to determining the bed of ariver for the purposes of the RMA.6 The correct legal test was regarded as the "bank tobank" test outlined in Kingdon v The Hutt River Board, which stands for theproposition that the bed of a river extends only from bank to bank, and morespecifically, is the area within the banks covered by water during the rainy season.7[11] The High Court held that the words "usual or non-flood" should be impliedand read into the RMA definition of "bed" before the words "fullest flow".8"Fullest flow" for the purpose of the definition was therefore to be regarded as:9 the river's fullest usual flow over a reasonable period of years of riveractivity cycles, and not including flood waters that would flow onto themargins and flood plain adjacent to the river.The application for leave to appeal[12] The Council applied for leave to bring a second appeal pursuant to s 303 ofthe Criminal Procedure Act 2011 concerning three discrete questions of law:(a) Did the High Court err in its assessment of the correct test fordetermining the extent of the riverbed in applying the definition of"bed" in s 2 of the RMA?(b) Did the High Court err in adding the phrase "usual or non-flood" intothe definition of "bed" in s 2 of the RMA by implication?5 At [7] and [67].6 Dewhirst Land Co Ltd v Canterbury Regional Council [2018] NZHC 3338, [2019] NZRMA 411.7 At [50]; referring to Kingdon v The Hutt River Board (1905) 25 NZLR 145 (SC).8 At [40].9 At [41].(c) Did the High Court err in concluding that the assessment of variousflow rates or return periods was an irrelevant consideration indetermining the extent of the riverbed?Each is said to involve matters of general or public importance. This Court has heldthat the threshold of general or public importance is met where a proposed appealraises an important question of law that has broad application beyond thecircumstances of the particular case.10 The respondents did not oppose the grant ofleave for the three questions of law.[13] We are satisfied that the issues raised in this appeal are of general and publicimportance. The correct approach to the RMA definition of "bed" will determine theextent of the area of land considered to be riverbed in rivers across New Zealand,particularly braided rivers. Moreover, the characterisation of land as riverbed has asignificant effect on how much land may be used by land owners, the RMA protectionswhich apply to it, and how it may be regulated under the RMA. We consider theproposed questions of law are likely to affect private individuals and entities, publicagencies (including all regional councils who have responsibility for the sustainablemanagement of rivers and their beds) and the New Zealand public as a whole.[14] For the above reasons we grant leave to the Council to appeal againstdeterminations of the High Court in respect of the three questions of law.The High Court judgment[15] Before addressing these questions, we will summarise the key findings fromthe judgment of Gendall J. The Judge held that in determining what is a riverbed, it isnecessary to ascertain the area covered by the river at its fullest flow and the locationof the banks of the river.11 The terms "bank" or "banks" are not defined in the RMA,but should be understood to mean the "raised border to a water feature that constrainsthe water's usual movement".12 Not all of a riverbed may contain water. On occasions10 R v Kuru [2015] NZCA 414 at [7].11 Dewhirst Land Co Ltd v Canterbury Regional Council, above n 6, at [19].12 At [24].there may even be no water at any point between the banks.13 However, not all waterthat flows from a river will be considered the river.14 Areas of land that are onlycovered by water in periods of flooding when the true banks of a river are overtoppedare the "margins" or "flood plains" that surround a river and are not the "bed" of ariver.15 Thus, a river's "fullest flow" must be something less than the point where itfloods.16[16] The Judge noted that the statutory term "fullest flow" is not defined inthe RMA.17 Read literally, the words would mean at its peak or at the largest floodlevel of the river.18 That could not be a correct interpretation as it would meanextending the banks of the river into adjoining land following significant floodevents.19 The Judge considered using a flood event to assist in the definition is"quite inappropriate".20 In ascertaining the "bed" of a river, the actual banks of a rivercourse, being the land alongside or sloping down to that river by way of visualinspection or otherwise, need to be established first.21[17] In the case of a braided river, the Judge held that the "bed" is the area of landwhich the waters of the river cover up to that bank when a significant flow eventarises.22 To say that the river bank needs to be located in such a position that it wouldnever or only very rarely be overtopped by flood waters is "simply nonsensical".Thus, a river's margin or floodplain should not form part of the "bed" of a river interms of a workable RMA definition.23 Accordingly the river "bed" should bedistinguished from the "margins" or "flood plains" that may surround a river.24[18] The Judge observed that, in considering where the river banks (with usualflow) would lie, a reasonable visual observation and a consideration of the river's13 At [26].14 At [27].15 At [36].16 At [28].17 At [30].18 At [32].19 At [31]–[32].20 At [33].21 At [34].22 At [34].23 At [34].24 At [36].natural character and the riverine qualities of the riverbed would be required as part ofthe exercise.25 In a braided river, the edges of the individual braids would not constitute"banks" of the river but rather the delineated banks where defined river stones or thelike slope up to land adjacent to the river margin (and possibly a river terrace or floodplain) would do so.26 Thus, the Judge found:27 the river bed is the area between the reasonably observable banks of a river.While not as clear as in other rivers, the sloping banks of a braided river maystill be found. The banks of the river are the border between land that may becovered by a significant but usual flow of the river, over a reasonableperiod of years of river activity cycles and land that is occasionally eroded orflooded by the river. The river bed is separate from the margin or flood plainwhich will often surround a river and the latter may, in some cases, be quiteextensive.[19] The Judge further held that:[50] The District Court did not apply the correct test here as it did not givesufficient weight to the true and observable location of the river's presentbanks and appeared to overlook the difference between a river "bed" and itsmargin or the adjoining flood plain. The purpose in the RMA of thedefinition of "bed" is to delineate between land that is, on the one hand, trulya river margin or flood plain and land, on the other, that is river bed.That purpose is achieved when one applies the final qualifying phrasethe legislature specifically selected to bring a limit to the words in s 2(modified by implication, for the reasons noted above) "at its fullest (usualand non-flood) flow" which does not overtop those banks. In determining the"bed" of a river, the long accepted "bank to bank" test outlined in cases suchas Kingdon v Hutt River Board is the correct test and is to be applied here.(Footnote omitted.)The first question of lawSubmissions of the Council[20] Mr Fowler QC, for the Council, asserted that the correct test was applied inthe second judgment in the District Court which held the RMA definition of "bed"means the space of land which the waters of the river cover with two conditions:28(a) first, at its fullest flow; and25 At [42].26 At [42].27 At [44].28 Canterbury Regional Council v Dewhirst, above n 2, at [61].(b) secondly, without overtopping its banks.[21] Mr Fowler submitted the two conditions are discrete in the sense that they arenot to be read together as a compound phrase. Rather, consideration of each issue willinform the other. It would not be proper to consider and determine either of these twoaspects of the definition in isolation of the other. In the case of a braided river likethe Selwyn, unlike a single thread river incised between two terraces, there will be anumber of topographical features which could be identified as banks and which mayshift over time. It would be improper and arbitrary to select one such feature anddesignate it as the bank without also considering the extent of the river's fullest flowand how regularly the various braids are flowing and how they may migrate.[22] Conversely, it would be incorrect to define the banks solely by reference tothe river's fullest flow. The definition should not be interpreted as effectivelystipulating that all of the area covered by water at the time of fullest flow must bewithin the banks and therefore that the banks to be identified must capture all of thatflow. The definition necessarily recognises that some of the area covered by waterwhen the river is at its fullest flow will not be riverbed, given that the river may haveovertopped its banks in places at the time of fullest flow. In this way the banks of theriver would still constrain what is properly regarded as riverbed, in that any landcovered by water at the river's fullest flow which is beyond the identified banks wouldnot qualify as riverbed.[23] For the purposes of the definition of "bed" in the case of a braided river(with varying flow rates and multiple topographic features which could be regardedas banks), a holistic assessment of relevant factors will determine the extent ofthe river's fullest flow and the position of the river's banks. Such an approach wasfavoured by Mr McCracken, who opined that a number of relevant factors should beconsidered, including the historical extent of the riverbed, flow modelling, LIDARanalysis,29 and the identification of observable topographical features which could becharacterised as banks. Under this approach, the identification of the banks is29 LIDAR (Light Detection and Ranging) is a form of surveying which can assist in identifyingtopographical features.informed by consideration of what might be regarded as the river's fullest flow (andvice versa).[24] Mr Fowler referred to the two parts to the definition of "bed". The first(in s 2(a)(i) of the RMA) relates to esplanade reserves, esplanade strips andsubdivisions in respect of which the words "fullest flow" are preceded by the word"annual". The second part (in s 2(a)(ii)) is for all other cases, in respect of whichthe words "fullest flow" are not so limited. Mr Fowler submitted it was illogical tosuggest that one interpretive approach should be applied to the definition of "bed" ins 2(a)(i) but that an entirely different interpretive approach should be applied tothe definition in s 2(a)(ii). The two definitions only differ by the insertion of a singleword ("annual").[25] With respect to the words "annual fullest flow", Mr Fowler submitted thephrase is equivalent to a river's "mean annual flood", a hydrological metric that canbe calculated by reference to flow rates and rainfall, and which is not affected ordictated by the position of a river's banks.30 Therefore, given that "annual fullestflow" is an independently quantified metric unaffected by the position of a river'sbanks, it is contradictory to suggest that a river's annual fullest flow is dictated orconstrained by the position of its banks.[26] In addition to urging the application of a holistic approach, Mr Fowler reliedon the purpose provision of the RMA31 and its accompanying provision emphasisingmatters of national importance32 to which we will refer later. These provisionsemphasise the need to mitigate adverse effects of external activities on theenvironment, including the inappropriate use and development of land in a way thatwould interfere with natural and physical resources. This is particularly relevant inthe present case where the unauthorised works were undertaken for the purpose ofdeveloping and protecting land which fell within the bed of the river for use as farmland.30 See Whitby Coastal Estates Ltd v Porirua City Council [2009] NZRMA 269 (EnvC) at [47].31 Resource Management Act, s 5.32 Section 6.[27] A second, and related, point is that the preservation of the natural character ofrivers is a matter of national importance in s 6(a) of the RMA. In the case of thenatural character of a braided river, it is dynamic, with multitudinous braids, includingthe main braid, shifting quite significantly over time (typically over decades). Thus, abraided river should be able to "behave" naturally and be preserved so that it can bepermitted to retain its natural character.[28] Mr Fowler also addressed the conclusion in the High Court judgment thatthe Council's concern regarding its ability to protect river environments wasunfounded.33 This was said by the Judge to be because local authorities can governgeneral river environments by the implementation of regional rules under s 30(1)(c)of the RMA or other regulatory controls such as bylaws.34 Mr Fowler submitted thata difficulty with this proposition is that it is premised on the fact that the relevant areaof land is not riverbed. This places a natural restriction on a local authority's abilityto regulate so as to protect riverine values.[29] Mr Fowler submitted a further problem arose with the High Court's approachin the case of braided rivers. As natural systems, the morphology, topography, bankposition, and flow regime of a particular river may vary greatly across the entire courseof that river and over a period of time.35 Any such variations will be compounded withdynamic braided rivers like the Selwyn River. Mr Fowler submitted thatthe High Court's interpretation would require a river's fullest flow to remain entirelywithin its identified banks. Applying this approach dictates that any breach of theidentified banks at any point along the river course would disqualify the particularflow rate from being considered the fullest flow. However, Mr Fowler accepted thatfor a river which follows a consistent course with well-defined banks (and flow ratesthat are closely aligned with those banks), such an approach may be appropriate.33 Dewhirst Land Co Ltd v Canterbury Regional Council, above n 6, at [48].34 At [48].35 Citing Jefferies v Wellington Regional Council [2012] NZEnvC 50.[30] Mr Fowler referred to a number of cases dealing with the definition of ariverbed,36 submitting the following propositions might be drawn from theseauthorities:(a) Where a river, as part of its natural processes, flows into a dried arm orbraid, that arm or braid is self-evidently part of the riverbed, even wherethe intermittence of that flow may be years apart.37(b) A river like the Selwyn, which many people may never have seenflowing, still has riverine values requiring protection as riverbed.This is to be distinguished from genuine floodwaters which may sheetacross adjoining land when a river has overtopped its banks.38(c) The cycles of river activity should be considered over a period ofdecades rather than in the short term.39 The riverbed is the area ofground over which the river flows and moves gravel over a period ofdecades. The riverbed encompasses the dynamic migration of a braidedriver over a period of time.40(d) The proper approach to defining the bed of a river or lake is to considerdifferent river flow rates or lake levels and to apply those in helping todetermine the proper location of the river's bed or the lake's margins.41[31] Finally, Mr Fowler submitted that the interpretation of "bed" adopted bythe High Court failed to consider all relevant factors, cannot be consistently andcoherently applied to the wider RMA definition of "bed", fails to achieve theunderlying purposes of the RMA, and does not recognise, preserve or protect thenatural character of braided rivers.36 Including Carruthers v Otago Regional Council [2013] NZHC 632, [2013] NZRMA 428;Canterbury Regional Council v Erralyn Farm Ltd DC Christchurch CRI-2011-003-1050,14 December 2012; and Jefferies v Wellington Regional Council [2013] NZHC 1059.37 Jefferies v Wellington Regional Council, above n 36, at [35].38 Carruthers v Otago Regional Council, above n 36, at [37]–[38].39 Canterbury Regional Council v Erralyn Farm Ltd, above n 36, at [29].40 At [46].41 At [46]; and Whitby Coastal Estates Ltd v Porirua City Council, above n 30, at [13].Respondents' submissions[32] The respondents support the approach of the High Court, namely, that todetermine the bed of a river, its banks must be identified first. The "banks" (a termleft undefined by the RMA) are the elevated features that define the side or border ofthe bed within which the river flows at its fullest in its ordinary or non-flood condition.If the banks were overtopped by water at its fullest flow, that water would be enteringthe margins and flood plain and would be leaving the bed of the river.[33] Mr Somerville QC for the respondents supported the use of the "bank to bank"test for determining the bed of a river as the correct test.42 The words "at its fullestflow" in the definition of "bed" were qualified by the final phrase"without overtopping its banks". As the High Court found, the definition of "bed"cannot be applied in a workable manner without first identifying where the relevantbanks are.43[34] Finally, Mr Somerville contended that the High Court was correct to find thedefinition of "bed" does not need to be strained or stretched in order to achieve whatthe Council considers to be the purposes of the RMA.44 The Council had available toit instruments such as the Regional Policy Statement, the Regional Land andWater Plan and district plans which provide regimes for addressing the naturalcharacter of rivers and their margins and natural hazards. Moreover, regional councilscan make regional rules concerning restrictions under s 13 of the RMA as part of theirs 30(1)(c) functions, which specifically include the avoidance or mitigation of naturalhazards.45 Regional councils can also rely on their powers under s 14 of the RMA toensure that the purpose of the RMA of protecting riverine values is achieved.46Applicable statutory provisions[35] The terms "bed" and "river" are both defined in s 2(1) of the RMA:bed means,—42 Applying Kingdon v The Hutt River Board, above n 7.43 Dewhirst Land Co Ltd v Canterbury Regional Council, above n 6, at [34].44 At [43].45 Resource Management Act, ss 6(h), 30(1)(c)(iv) and 68.46 Section 14(2) provides that no person may divert any water without express permission.(a) in relation to any river—(i) for the purposes of esplanade reserves, esplanade strips, andsubdivision, the space of land which the waters of the rivercover at its annual fullest flow without overtopping its banks:(ii) in all other cases, the space of land which the waters of theriver cover at its fullest flow without overtopping its banks;and(b) in relation to any lake, except a lake controlled by artificial means,—(i) for the purposes of esplanade reserves, esplanade strips, andsubdivision, the space of land which the waters of the lakecover at its annual highest level without exceeding its margin:(ii) in all other cases, the space of land which the waters of thelake cover at its highest level without exceeding its margin;andriver means a continually or intermittently flowing body of fresh water; andincludes a stream and modified watercourse; but does not include any artificialwatercourse (including an irrigation canal, water supply race, canal for thesupply of water for electricity power generation, and farm drainage canal).[36] Of relevance is the fact that the terms "fullest flow" and "banks", used inthe definition of "bed" in s 2(1), are not defined in the RMA.[37] Sections 5 and 6 of the RMA are also relevant:5 Purpose(1) The purpose of this Act is to promote the sustainable management ofnatural and physical resources.(2) In this Act, sustainable management means managing the use,development, and protection of natural and physical resources in away, or at a rate, which enables people and communities to providefor their social, economic, and cultural well-being and for their healthand safety while—(a) sustaining the potential of natural and physical resources(excluding minerals) to meet the reasonably foreseeable needsof future generations; and(b) safeguarding the life-supporting capacity of air, water, soil,and ecosystems; and(c) avoiding, remedying, or mitigating any adverse effects ofactivities on the environment.6 Matters of national importanceIn achieving the purpose of this Act, all persons exercising functionsand powers under it, in relation to managing the use, development,and protection of natural and physical resources, shall recognise andprovide for the following matters of national importance:(a) the preservation of the natural character of the coastalenvironment (including the coastal marine area), wetlands,and lakes and rivers and their margins, and the protection ofthem from inappropriate subdivision, use, and development:(b) the protection of outstanding natural features and landscapesfrom inappropriate subdivision, use, and development:(c) the protection of areas of significant indigenous vegetationand significant habitats of indigenous fauna:(d) the maintenance and enhancement of public access to andalong the coastal marine area, lakes, and rivers:(e) the relationship of Maori and their culture and traditions withtheir ancestral lands, water, sites, waahi tapu, and othertaonga:(f) the protection of historic heritage from inappropriatesubdivision, use, and development:(g) the protection of protected customary rights:(h) the management of significant risks from natural hazards.[38] In the present context, the terms "bed" and "river" are used in connection withstatutory restrictions on the uses of beds of rivers. Set out below is the relevant portionof s 13 of the RMA:River and lake beds13 Restriction on certain uses of beds of lakes and rivers(1) No person may, in relation to the bed of any lake or river,—(a) use, erect, reconstruct, place, alter, extend, remove, ordemolish any structure or part of any structure in, on, under,or over the bed; or(b) excavate, drill, tunnel, or otherwise disturb the bed; or(c) introduce or plant any plant or any part of any plant (whetherexotic or indigenous) in, on, or under the bed; or(d) deposit any substance in, on, or under the bed; or(e) reclaim or drain the bed—unless expressly allowed by a national environmental standard, a rulein a regional plan as well as a rule in a proposed regional plan for thesame region (if there is one), or a resource consent.(2) No person may do an activity described in subsection (2A) in amanner that contravenes a national environmental standard or aregional rule unless the activity—(a) is expressly allowed by a resource consent; or(b) is an activity allowed by section 20A.(2A) The activities are—(a) to enter onto or pass across the bed of a lake or river:(b) to damage, destroy, disturb, or remove a plant or a part of aplant, whether exotic or indigenous, in, on, or under the bedof a lake or river:(c) to damage, destroy, disturb, or remove the habitats of plantsor parts of plants, whether exotic or indigenous, in, on, orunder the bed of a lake or river:(d) to damage, destroy, disturb, or remove the habitats of animalsin, on, or under the bed of a lake or river.Our analysis[39] The Council's interpretation of a bed of a river focuses first and foremost onthe area covered by the river at its fullest flow and less upon the location of the banksof the river. Hence, the Council's starting point, relying on the words "at its fullestflow", is the flow of the river. Such an approach would permit the Council's expert,for example, to reply upon a one in 50-year flood as a proper means of identifyingwhere the boundary of the bed might be.47 We say at the outset we see significantdifficulties with this interpretation, the details of which we explain below.[40] Although the term "bed" in relation to any river is found in s 2 of the RMA,enacted in 1991, we consider a proper starting point is to trace the term back to itsstatutory origins and earlier. The word appeared in a statutory context in 1903 whenthe Coal-mines Act 1891 was amended by s 14 of the Coal-mines Act Amendment Act1903. The effect was to confirm Crown ownership of navigable riverbeds when title47 A point referred to by Gendall J in Dewhirst Land Co Ltd v Canterbury Regional Council, aboven 6, at [53].to the bed had never been alienated by the Crown. The 1903 Amendment thus defined"Bed" as meaning "the space of land which the waters of the river cover at its fullestflow without overflowing its banks".48 The Conservation Act 1987, the CrownMinerals Act 1991, and the Overseas Investment Act 2005 all adopted a similardefinition of "bed" in relation to a river.49 In each case the words "overflowing itsbanks" were replaced with the words "overtopping the banks".[41] Where an Act of Parliament uses a phrase such as "bed in relation to anyriver", a pertinent question is whether it had an established meaning at common law.This may provide a valuable insight into Parliament's intention when the phrase wasused in 1903 and subsequently.50 We therefore examine what the common law had tosay about the definition of "bed" in relation to a river.The common law[42] The meaning of the "bed" of a river cannot realistically be separated from themeaning of a river and its banks. This is apparent from the discussion of the conceptof a river in the 1887 case of Mayor of Rangiora v Ashley Road District.51 The caseconcerned whether a traffic bridge was a bridge for the purposes of the Public WorksAct 1882 Amendment Act 1884. The issue was whether the bridge structure neededto extend from bank to bank across the river. In considering the meaning of a river,Johnston J found the existence of "banks" to be a more or less essential ingredient inthe proper definition of a river:52The word "bank" implies an acclivity or elevation of land above the level ofthe adjacent land or water, and a river-bank is a boundary which is of sufficientelevation to restrain the water from flowing over the adjacent land underordinary circumstances. In many, or most, of the rivers of this colony, weknow that the banks are often indefinite and liable to constant changes, as arethe currents in their beds, 48 Coal-mines Act Amendment 1903, s 14(2).49 Conservation Act 1987, s 2(1); Crown Minerals Act 1991, s 2(1); and Overseas Investment Act2005, s 6(1).50 See Diggory Bailey and Luke Norbury (eds) Bennion on Statutory Interpretation (7th ed,LexisNexis, London, 2017) at [25.1]–[25.4].51 Mayor of Rangiora v Ashley Road District (1887) 6 NZLR 119 (SC).52 At 123.[43] A leading common law text on rivers is Houck's A Treatise on the Law ofNavigable Rivers.53 The concept of a river is defined as comprising a bed, a shore andbanks.54 Houck's definition is elaborated thus:55The bank is the outermost part of the bed in which the river naturally flows.In the words of the Digest, "that is considered to be bank which contains theriver when fullest." The bed is covered by the river, and is the space subjacentto the river, through which it flows. The shore, or beach of a river may bedefined as that part of the river-bed lying between the top of the bank and thatpart of the bed where the water actually flows, and which, as the water risesor falls, is land or river.[44] More specifically, Houck defined the bank of a river by reference to its bed:56The bank is then, "the outermost part of the bed in which the river naturallyflows at its fullest." The bank may be thus rightly defined, as that line orridge of earth which contains the river, holding the natural direction of itscourse. But if at any time, either from rains, the sea, or any other cause, it hasoverflowed for a time that line, it does not by such overflow change its banks.Nobody has said that the Nile, which by its overflow covers Egypt, changesor enlarges its banks; for when it has returned to its usual height, the banks ofits bed are to be secured. Thus it appears, that overflowing does not alter thebank of a river; because that overflow is for a time only, while the natural flowis more or less constant.[45] Houck defined the bed of a river as follows:57The bed is covered by the river, and is the space subjacent to the river, overwhich it flows, between the banks. It is the space between the banks occupiedby the river at its fullest flow. The bank is the border of the bed, withinwhich bed the river flows when in its fullest state naturally; that is to say,when not temporarily overflowed by extraordinary rains. Then it follows, thatall water contained in the river's bed, between the two banks, or theirhigh-water line, is river.[46] For the above propositions Houck cited the United States Supreme Court caseof Howard v Ingersoll,58 from which we quote only a short passage from the opinionof Justice Curtis who stated:5953 Louis Houck A Treatise on the Law of Navigable Rivers, (Little, Brown & Co, Boston, 1868).54 At 3.55 At 3 (footnotes omitted).56 At 5 (footnotes omitted).57 At 5–6 (footnotes omitted).58 For an early discussion of the nature and scope of the bed or channel of a river in England, see R vThe Inhabitants of Oxfordshire (1830) 1 B & Ad 289, 109 ER 794 (KB).59 Howard v Ingersoll 54 US 381 (1851) at 427–428. the banks of a river are those elevations of land which confine the waterswhen they rise out of the bed; and the bed is that soil so usually covered bywater as to be distinguishable from the banks, by the character of the soil, orvegetation, or both, produced by the common presence and action of flowingwater. But neither the line of ordinary high-water mark, nor of ordinarylow-water mark, nor of a middle stage of water, can be assumed as the linedividing the bed from the banks. This line is to be found by examining thebed and banks, and ascertaining where the presence and action of water are socommon and usual, and so long continued in all ordinary years, as to markupon the soil of the bed a character distinct from that of the banks, in respectto vegetation, as well as in respect to the nature of the soil itself. But inall cases the bed of a river is a natural object, and is to be sought for, not merelyby the application of any abstract rules, but as other natural objects are soughtfor and found, by the distinctive appearances they present; the banks beingfast land, on which vegetation, appropriate to such land in the particularlocality, grows wherever the bank is not too steep to permit such growth, andthe bed being soil of a different character and having no vegetation, or onlysuch as exists when commonly submerged in water.[47] This definition was applied by Romer J in the Chancery Division in Hindson vAshby — a case concerning changes to the bed of the River Thames.60 The principlesadopted by the Judge were later endorsed by the English Court of Appeal, althoughthe Judges reached a different conclusion on the facts.61[48] The Court of Appeal of New Zealand in 1893 considered the meaning of"river" or "watercourse" in Piripi Te Maari v Matthews.62 The case concerned thewatercourse or channel leading from Wairarapa Lake to the sea. During a considerablepart of each year the lake had no outlet and the channel was dry on account of theexistence of a sand and shingle bar. A majority found that the channel was neverthelessa natural watercourse for the purposes of s 184 of the Public Works Act 1882.63The majority judgment was given by Conolly J, who stated that, even if rivers orwatercourses are dry for many months at a time, no-one could question "that they arenot the less watercourses even when dry".64[49] The importance of the banks of a river, when determining the nature and scopeof the river's bed, was discussed in the Scottish case of Menzies v Marquis ofBreadalbane.65 The dispute was between riparian owners who also owned the bed60 Hindson v Ashby [1896] 1 Ch 78 (Ch) at 84–85.61 Hindson v Ashby [1896] 2 Ch 1 (CA) per Lindley LJ at 14.62 Piripi Te Maari v Matthews (1893) 12 NZLR 13 (CA) at 22.63 Comprising Prendergast CJ and Richmond, Denniston and Conolly JJ.64 At 22.65 Menzies v Marquis of Breadalbane 1901 SLR 35 (IH).(alveus) of the river Tay up to the midpoint (medium filum). The question was wherethat midpoint was.66 Lord Trayner found that at the point in question the river flowedbetween well-defined banks.67 Moreover, on the facts, when ordinarily full the rivercovered the whole of the bed from bank to bank.68 This was so even though there wasmore water on the one side than on the other because the channel was deeper on oneside.69 Lord Trayner held that "where the water of a river in its ordinary conditioncovers the alveus from bank to bank it is the centre of the alveus between the banksthat is the medium filum, and consequently the boundary of the properties on theopposite banks".70[50] The above authorities represent a useful, albeit non-exhaustive, summary ofthe common law principles in relation to the meaning of the terms "river", "bed" and"bank" as at 1903.71 The legislative drafters contemplating the amendment tothe Coal-mines Act at that time would have had an awareness of such establishedmeanings of these relevant terms when enacting the amendment to that Act.Summary of common law principles pertaining to rivers and beds[51] To assist the analysis in this case, the following principles emerge from theabove common law authorities:(a) The description of a river or watercourse includes as essential featuresthe channel (or bed) and its banks.(b) The bed comprises the space between the banks occupied by the riverat its fullest flow.66 At 37.67 At 37.68 At 37–38.69 At 38.70 At 38.71 We have not overlooked the English Court of Appeal decision on the meaning of "bed" of a riverin Conservators of the River Thames v Smeed, Dean & Co [1897] 2 QB 334 (CA) in whichA L Smith LJ cited an earlier American case of The State of Alabama v The State of Georgia 64US 505 (1859), a case which was also cited in the Court of Appeal in Hindson v Ashby, above n 61,at 25–26.(c) Ascertaining the bed in a given case will require consideration of allrelevant geographical, meteorological and hydrological features suchas banks, channels, shores, seasonal flows, as well as unseasonable wetweather events which produce a flood where the water overflows thebanks and spreads into the surrounding areas.(d) The bed of a river is not limited to the portion between the banksthrough which the water flows only in dry weather. Equally, though ariver or watercourse is dry for part (even the greater part) of the year, itis nonetheless a river or watercourse.(e) The bank of a river is the outermost part of the bed and comprises anacclivity or elevation of land above the level of the adjacent land orwater, which creates a boundary sufficient to prevent the water fromflowing into the neighbouring land.(f) The banks of some rivers may often be indistinct or indefinite and liableto constant changes, as are the waters or currents in their beds.(g) All the water contained in the riverbed, between the two banks, or theirhigh-water line on the shore, is river.(h) Ascertaining the dividing line between the banks and the bed cannot bedetermined by reference only to the ordinary high water mark or theordinary low water mark. The task requires examining the banks andthe bed and finding where the presence and action of the water arecommon or usual as apply in ordinary years. Also relevant are thenature of the soils of the bed and the banks and the surroundingvegetation.(i) The bed of a river is a natural object to be determined not by abstractrules, but by the distinctive appearances they present, particularly inrespect of the banks and their soils and vegetation.Development of New Zealand law on rivers, banks and beds[52] A Full Court of the Supreme Court considered the words "river", "banks" and"bed" in 1904–1905 in Kingdon v The Hutt River Board.72 The Chief Justice asPresident of the Compensation Court stated a case which directly raised the meaningof these three words as they applied to the Hutt River Board's control overthe Hutt River. In the course of argument counsel referred to many of the leadingtreatises on rivers, as well as the authoritative cases from the United States of Americaand the United Kingdom. One of the questions in the case stated required the Courtto define what in law constitutes the "bed" of the Hutt River.73 Despite the fact therewas a definition of "bed" in relation to a river in the Coal-mines Act, the Court turnedfor guidance to the common law. The Hutt River had defined banks but the flow ofwater between such banks was irregular.74 During the dry months the water flowed ina small channel while in the wet weather the flow was greatly increased. At this timethe river was said to flow from bank to bank in circumstances described as "ordinaryfreshes".75 In very wet weather the river would be "in flood" at which time it wouldoverflow its banks.76[53] The judgment of the Full Court was given by Stout CJ who referred to thedecision in Piripi Te Maari v Matthews — which considered the question of whatcomprised the bed of a watercourse.77 Reference was also made to English andAmerican authorities cited by counsel to support the proposition that the bed of a riveris not limited to that portion between the banks through which the water flows in dryweather. The Chief Justice then discussed the meaning of "freshes" inthe New Zealand context, stating:78A "fresh" in a New Zealand river, when the water is confined within the banksof the river, is nothing more than what may be termed the ordinary conditionof the river during the rainy season, and the evidence in the present case is thatduring the rainy season the water in the river, flowing in much greater quantitythan in the dry season, frequently reaches the banks on either side.In exceptional instances, happening once in every two or three years, when72 Kingdon v The Hutt River Board, above n 7.73 At 149.74 At 156.75 At 156–157.76 At 157.77 Piripi Te Maari v Matthews, above n 62.78 Kingdon v The Hutt River Board, above n 7, at 157–158.the rainfall has been long continued and of great severity, the water in the riverbecomes a "flood", and overflows its banks and spreads over the adjacentcountry. But in the ordinary rainy season the water is confined within thebanks, and as it frequently extends from bank to bank during such season, andreaches the bank on the claimant's land, we are of opinion that the shinglefrom bank to bank is within what may properly be called the "bed" of the river,although the water of the river does not in the dry weather ordinarily flow oversuch shingle.[54] Mr Fowler was critical of Gendall J for adopting the bank to bank test outlinedin Kingdon. He argued that the case concerned provisions of the River Boards Act1884 which did not define either "river" or "bed". Nor was it necessary for the Courtin Kingdon to consider the concept of "fullest flow" or any other aspect of the RMAdefinition of bed. The case involved an entirely different statutory context.[55] We consider such criticism of Kingdon, and of Gendall J for following it, ismisplaced. It is true that it concerned a different statutory context, but the words"river" and "bed" still needed to be defined. A careful reading of the case reveals thatthe Court reviewed the principles emerging from the treatises on rivers and thecommon law cases. What the Court in Kingdon did was to apply such principles tothe issues for determination involving the nature and scope of the Hutt River and itsbed. Further, Kingdon was decided at a time when "fullest flow" was a concept thatwas already established in common law and in statute (see [40]–[45] and [51(b)]above). It follows that we see no error on the part of Gendall J for relying on Kingdonas part of his reasoning.[56] Gendall J also referred to a decision of Fogarty J in the High Court inCarruthers v Otago Regional Council which discussed the definition of bed in s 2 ofthe RMA.79 The case concerned a prosecution for disturbance of a riverbed contraryto s 13(1) of the RMA. On appeal the question was whether the actions ofthe appellant, Mr Carruthers, amounted to a breach of this provision and in particularwhat on the facts amounted to a "river" as opposed to an artificial watercourse.80[57] The Judge accepted that, in the context of a s 13(1) prosecution, there was meritin protecting riverbeds subject to intermittent flows.81 He added:79 Carruthers v Otago Regional Council, above n 36.80 At [5].81 At [37].[37] There are, of course, many creeks in New Zealand that can bereasonably found to qualify as "rivers" because they do flow intermittently.Many side creeks of high country rivers fall into that category. They do havea bed which regularly fills and, in extreme weather environments, floods, asin extends beyond its natural bed. You can say that they are streams, eventhough in late summer and autumn they may have dried up, at least visiblyfrom the surface. It is important that the beds of those side creeks not bedisturbed, as they fulfil an important function of clearing water from acatchment without doing damage to the bush on either side of the creeks.[38] Similarly, there are riverbeds in pastoral areas which have the samefunction. One thinks immediately of the Selwyn River in Canterbury.Many Cantabrians have never seen water in that riverbed. It does not meanthat the riverbed does not have a function and a need for it to be preserved.[58] The appeal also raised an issue of whether, in times of major storms, the sheetsof water which spread across nearby alluvial fans could comprise part of the river.Fogarty J considered it important not to treat such sheets of water as a river.82Referring to the definition of "bed" in s 2 of the RMA, the Judge said this:[41] This definition makes it clear that Parliament never intended tosuggest that floodwaters, or flows that follow only from major storms, fellwithin the definition of river. On the contrary. In that sense, one cannotfully understand the definition of river in s 2 without taking into account thedefinition of bed.[59] We consider Gendall J was right to apply the observations of Fogarty J.The normal or usual flow of a river (putting aside the concept of "fullest flow" towhich we will return) is plainly quite different from when the river floods. Such anobvious distinction was well understood, and applied, in the common law cases.Bed of a river — statutory context[60] For present purposes, the concept of a riverbed falls to be considered in relationto the set of restrictions on use stipulated in s 13 of the RMA. That suggests the bedof the relevant river is seen as a natural object, the uses of which the law has seen fitto regulate and protect, for example, by controlling its use for structures, excavation,drilling, tunnelling or similar disturbances and the introduction of any plant matter.The restrictions extend to depositing any substance in, on, or under the bed, reclaimingor draining the bed and to a range of activities which could damage, destroy or disturbthe bed as proscribed in s 13(2A).82 At [40].[61] The statutory definition of "bed" in s 2 of the RMA applies in relation to anyriver, emphasising the universal nature of its application for the purposes of the RMA.No doubt the definition of "bed" will, in many cases, be of straightforward applicationon the facts. This will not always be so, especially when, as was said in Mayor ofRangiora, the banks are "indefinite and liable to constant changes, as are the currentsin their beds".83[62] The concept of a riverbed being a natural object is consistent with the commonlaw principle noted at [51(i)] above. This approach is underscored by the fact that"bed" is defined in s 2(a)(ii) of the RMA as the space of land which the waters ofthe river cover. In order to ascertain the nature and scope of the space of land inquestion, one might refer first to the RMA definition of "river". However, suchdefinition is of limited assistance in that it refers only to a "body of fresh water" whichis flowing "continually or intermittently". The definition is extended to include astream or modified (but not an artificial) watercourse. One derives no further guidanceas to the geographical, hydrological or riverine features which might inform the natureof the space of land.[63] For the purposes of the RMA definition of "bed" in s 2(a)(ii), a "space of land"will have a physical or geographical boundary. In the case of a river, this will compriseits banks. It is hardly surprising then to find in the early common law cases statementssuch as banks are "a more or less essential ingredient in the proper definition of a'river'".84 The fact that the banks may be difficult to discern or may overflow fromtime to time creating a flood does not change this fundamental character of a river —its banks are a vital feature. This proposition is consistent with the early treatises onthe issue.[64] Before leaving the consideration of the physical or geographical boundary ofa riverbed we refer briefly to the related physical features of "margins" and "floodplains", which Gendall J referred to in his judgment.85 As to flood plains, he describedthese (accurately in our view) as "those areas of land that are only covered by water83 Mayor of Rangiora v Ashley Road District, above n 51, at 123. See also Canterbury RegionalCouncil v Erralyn Farm Ltd, above n 36, at [33].84 Mayor of Rangiora v Ashley Road District, above n 51, at 123.85 Dewhirst Land Co Ltd v Canterbury Regional Council, above n 6, at [36].in periods of flooding when the true banks of a river are overtopped".86 No furtherelaboration is required.[65] The term "margins" requires more attention. First, it is not defined inthe RMA. Second, it is plainly important, as it is referred to in s 6(a) in which the"preservation of the natural character of lakes and rivers and their margins" isstipulated as a matter of national importance. There is some discussion of the term inHigh Country Rosehip Orchards Ltd v Mackenzie District Council, also in the contextof s 6(a) of the RMA.87 There the Court expressed the obiter view that the terms"margins" and "banks" are not synonymous:88"Margin" in section 6(a) may have a different meaning from its use in section230(3) where it is confined to being used of lakes only as the equivalent ofthe "banks" of a river. In contrast the "margins" of lakes and rivers must be awider term. Margins are likely to be areas beyond the wave action of a lakeor extending away from the banks of a river for, depending on topography andother factors, at least 20–50 metres and sometimes more.[66] Finally, we note that the term "margin" is referred to in the CanterburyRegional Council Policy Statement as follows:89 the land immediately adjacent to the bed of a river, wetland, lake or estuarywhich is likely to be affected by a high water table, flooding, fluvial erosion,or sediment deposition, and often contains distinctive vegetation. The size ofthe margin will vary according to local site factors but may extend to the limitsdemarcated by natural river terraces and constructed stop banks.[67] Given that the margins of the river Selwyn are not directly relevant to the firstquestion of law, we say nothing further about the meaning of "margins".[68] The next feature of the RMA definition of "bed" in s 2(a)(ii) we address is thereference to the word "banks" in relation to the phrase "fullest flow withoutovertopping its banks". We consider this reference assumes the importance andrelevance of banks in understanding the space of land which comprises the bed of any86 At [36].87 High Country Rosehip Orchards Ltd v Mackenzie District Council [2011] NZEnvC 387 at [138]–[140].88 At [140].89 Environment Canterbury Regional Policy Statement (July 2019) at [10.1.2].river. The common law cases, as well as the early treatises, likewise assumed therelevance of banks. The statutory context does not suggest a different approach.[69] As we have noted, the term "banks" is not defined in the RMA. However, inany given case, the banks are likely to be important in identifying the outer extremitiesof the riverbed, as established by principle (b) at [51] above. The banks will also beimportant in differentiating geographically between the margins of a river or itsflood plains, a proposition accepted by Gendall J. It is for this reason that an importantfactual question in determining the "bed" of a river in a particular case will be toidentify the banks of that river in question which will set the outer limit of itsfullest flow.[70] Identifying the distinction between the bed of a river, its margins and floodplains, is a suitable point at which to respond to the Council's submission that thedefinition of "bed" as found by the High Court "remains premised on the relevant areaof land not being river bed" with the result this restricts the local authority's ability toregulate river margins and flood plains so as to protect riverine values.[71] We reject this submission. First, under s 14 of the RMA, the Council mayenforce restrictions relating to water. The ability of the Council to control thediversion of water is not dependent on the area in question being treated as riverbed.Moreover, there are other non-RMA regulatory controls available to the Council suchas its Flood Protection and Drainage Bylaw 2013 which facilitates management ofrisks from flooding in this particular flood plain.90[72] We agree with the submission of Mr Somerville that there is nothing to preventthe Council from controlling the use of land (which includes the flood plain), to fulfilthe requirement to avoid or mitigate the effect of natural hazards. The term"natural hazard" is defined as s 2 of the RMA to mean any atmospheric or earth orwater related occurrence, including flooding, the action of which adversely affects ormay adversely affect the environment.90 Canterbury Regional Council Flood Protection and Drainage Bylaw 2013 (January 2019).[73] We also accept Mr Somerville's submission that the definition of "bed" shouldnot be given a different meaning to accommodate the fact that a regional council didnot develop sufficient regional plan provisions to address such natural hazardcontingencies. It follows that Gendall J was correct to find the Council's statedconcern about its ability to protect river environments was unfounded.91Local authorities have the power to preserve the natural character of rivers and theirmargins and to manage the effects of natural hazards on land use within the margins,should they need to use it.92[74] We return to focus on the relevance of the water coverage aspect of thedefinition of "bed". This component of the definition is a further element in definingthe physical or geographical extremities of the riverbed. Houck defined the bed of ariver as "the space between the banks occupied by the river at its fullest flow" — seeprinciple (b) at [51] above.93 It is not surprising that the statutory definition of "bed"identified the water coverage of the river as at its fullest flow. If the legislators had inmind the common law approach, this suggests the water coverage element wasintended to be an additional factor to the identification of the banks of the river.[75] We recognise the wisdom of the statutory definition's choice of a multiplemeasure for determining the outer limits of the bed of a river. Given that the definitionapplies to all rivers, including as Mr Fowler emphasised, braided rivers such asthe Selwyn, in varying geographic, topographical, hydrological, climatic, seasonable,geological and vegetational conditions, the use of additional means of assessingthe bed is understandable.[76] But what the choice of the "fullest flow" measure does not do is elevate thisfactor to the prime (let alone the sole) criterion for measurement. We see no suggestionwithin the words of the definition driving this interpretation. Neither does the statutorycontext require such an outcome. We agree with the observations of Fogarty J inCarruthers that Parliament never intended that floodwaters or flows following only91 Dewhirst Land Co Ltd v Canterbury Regional Council, above n 6, at [47]–[48].92 At [48].93 Houck, above n 53, at 5–6.from major storms fell within the RMA definition of "river" when assessing ariverbed.94[77] It follows that we also agree with the view of Gendall J that a river's "fullestflow" for the purposes of the definition of "bed" must be something less than the pointwhere it floods.95[78] It was at this point of his analysis that Gendall J invoked the principle (correctlyin our view) that the bed of a river comprises those lands covered by water during theordinary rainy season, but contained within the banks of the river and extending frombank to bank.96 Such an approach is entirely consistent with the common law and theprinciples set out in the treatises discussed above.[79] Accordingly, we consider that the determination of the "bed" of a river, asdefined in s 2(a)(ii), will depend not only on the position of the banks of the river, butalso on the water coverage measure as determined by the river's fullest flow whichoccurs within those banks. This latter criterion is qualified by the words "withoutovertopping its banks". This qualifying term serves to exclude flows or inundationsarising from major storms where the water extends temporarily beyond the banks.[80] We would not exclude from consideration in particular cases those featuresidentified in principles (c), (d) and (h) in [51] above. These factors are all endorsedby the common law cases and we see no basis for excluding them in the wording ofthe s 2 definition of "bed".[81] In addressing the term "fullest flow", we have up to this point been applyingthe definition of "bed" in s 2(a)(ii) of the RMA as it relates to the present case. Yet, forthe purpose of esplanade reserves or esplanade strips in the definition of "bed" ins 2(a)(i), the different terminology of "annual fullest flow" applies. As Mr Fowlerrelied on the difference in wording between the two definitions as being significantwe will now address its relevance.94 Carruthers v Otago Regional Council, above n 36, at [41].95 Dewhirst Land Co Ltd v Canterbury Regional Council, above n 6, at [28].96 At [29].[82] Section 2(a)(i) of the definition was introduced by the Resource ManagementAmendment Act 1993. At the same time the concept of esplanade strips was alsointroduced into the RMA.97 It is likely that the new subparagraph was inserted torecognise that delineation of the riverward edge of esplanade strips need not be thesame as that which determines the area which requires protection under s 13 ofthe RMA.98 The 1993 Amendment left the definition of "bed" in what became s2(a)(ii) unchanged, apart from changing the words "the banks" to "its banks".[83] In view of this amendment, and in reliance on a decision ofthe Environment Court in Whitby Coastal Estates Ltd v Porirua City Council,99Mr Fowler submitted the position of the banks did not qualify or dictate the extent ofthe annual fullest flow. He submitted that the same approach should logically applyin s 2(a)(ii).[84] In Whitby, the term "annual fullest flow" was found to be the equivalent of ariver's mean annual flood.100 This involves a hydrological metric that can be definitelycalculated by reference to flow rates and rainfall, and which is not affected or dictatedby the position of a river's banks.101 Counsel also relied on an observation in Whitbythat the fullest flow of a river would generally be larger than its annual fullest flow.102[85] The argument advanced by the Council is that to give effect to the term "annualfullest flow", the phrase "without overtopping its banks" must qualify or increase theextent of the land covered by water at the annual fullest flow, rather than simplyconstraining the extent of the annual fullest flow itself. Thus, the only possible andworkable interpretation for the definition of "bed" relating to esplanades in s 2(a)(i) isthat postulated by the Council, in which the physical feature of the banks are of limitedor of no relevance.97 Provision for esplanade reserves was included in the original RMA.98 Esplanade strips and esplanade reserves are vested to enable public access to or along any river:s 229(b) of the RMA. They give effect to s 6(d), which provides for "the maintenance andenhancement of public access to and along rivers" as a matter of national importance.99 Whitby Coastal Estates Ltd v Porirua City Council, above n 30.100 At [54]–[56] and [76].101 At [47].102 At [17].[86] We do not accept that the significance of a river's banks in ascertaining thespace of land which the waters of the river cover (as stated in the RMA definition of"bed" in s 2(a)(ii)) should be cast aside by a legislative sidewind in the enactment ofs 2(a)(i). This cannot have been the legislative intention, particularly when thedefinition of "bed" in all other cases as covered by s 2(a)(ii) was left largelyunchanged.[87] Whitby is of limited direct relevance in the present case. However, we dealwith it further because it was relied upon by the Council. The case concerned whetherthe Duck Creek waterway in Porirua qualified for the taking of esplanade reserves forthe purposes of s 230 of the RMA. The Court was required to determine whetherthe Duck Creek was a river with a bed width of three metres or more.103 If anesplanade was required, it was necessary to consider the meaning of the phrase "alongthe bank of any river" in s 230(3) of the RMA and to identify the position on the bankfrom which the esplanade reserve should be set off.[88] Accordingly, the Court's focus was on the setting off of an esplanade reserveand the application of the definition of "bed" in s 2(a)(i), which required considerationof the words "annual fullest flow". The Court was concerned about the nature of theevidence required to enable it to determine the nature and scope of the bed ofthe Duck Creek. In particular, the question was whether expert hydrological evidencemight be of assistance.[89] The Court's attention was drawn to the fact that, when the amended definitionof "bed" was introduced into the RMA, officials from the Ministry forthe Environment informed the Select Committee that it had not been intended thathydrological data would be required, as this was only available for major rivers.They further advised that vegetation would indicate approximately where the annualfullest flow is for most rivers.104103 At [10].104 At [39].[90] It was unsurprising therefore that the Court accepted that hydrological data willnot always be necessary to determine the width of the bed of a river.105 Consistent withthe principles distilled from the common law, other factors may be sufficient.[91] Depending on the nature of the particular riverbed, the Court acknowledgedthat hydrological data may illuminate certain factual aspects which are required to beconsidered.106 The extent of the waters of the river will be important (as the RMAdefinition provides) and hydrology may assist in describing such extent in space andover time.107[92] In Whitby, the Court relied on hydrological evidence to find (as we have noted)that the annual fullest flow should be determined by reference to the Mean AnnualFlood (the MAF).108 On the facts, such a measure extended beyond the incisedchannel of the Duck Creek.109 However, the Court was careful to recognise that,where the MAF spills out into floodplain, the bank is the point where the spilloccurs.110 The bed width at this point is the width measured from bank to bank.This conclusion recognises, as we have earlier discussed, that the bed of a river is tobe distinguished from its flood plains.[93] Accordingly, in the context of the application of the RMA definition of "bed"in s 2(a)(i) (dealing with esplanade reserves), where the annual fullest flow of waterin a river extends beyond the banks, those banks will be a relevant (and limiting) factor.The fact that the banks are overtopped by a flood event does not change the outerlimits of the river. Whitby provides no support for the Council's interpretation.Even if (contrary to our interpretation) the wording of s 2(a)(i) supported the Council'sapproach, there is no warrant to apply the same approach to the different circumstancesand different wording in s 2(a)(ii).[94] We now address the Council's submission that the definition of "bed" ins 2(a)(ii) should be given a purposive interpretation. Such an approach was advanced105 At [46].106 At [46].107 At [46].108 At [47]–[48].109 At [50]–[52].110 At [53].by Mr Fowler to support the Council's primary focus on the fullest flow of a river asa defining feature of the geographic extent of the land space of its bed.Counsel invoked ss 5 and 6 of the RMA as driving an interpretation which wouldmitigate the adverse effects of external activities on the environment, including theinappropriate use and development of land in a way that would interfere with naturaland physical resources.[95] We accept the force of these submissions. Moreover, we have not overlookedthe various matters of national importance listed in (a)–(h) of s 6 of the RMA.Yet when it came to defining the term "bed" in relation to a river, the RMA definedthe term in a manner that was apt for the statutory context, including the companiondefinition of "river" and the common law from which these and related terms arederived.[96] We also recognise the force in Mr Somerville's submission that the Council'sexpansive definition of "bed", to meet a purposive interpretation, carries real practicaldifficulties. As Gendall J noted, much land currently used for various purposes, suchas general farming, roads or even buildings, could be included within the scope of sucha definition.111 We agree with the Judge's observation that: 112There are obvious difficulties that face a landowner or controlling body as towhether land adjacent to a river (especially a braided river) that is usually dryis still classified as river "bed" or otherwise, and therefore whether that landcould be used for other purposes or not. Given the limited activities that maybe permitted on a river bed, an expansive definition of what is truly "bed",coupled with this uncertainty of just how far that bed might extend, cannothave been Parliament's intention.[97] We have already discussed the other means which the Council, and other localauthorities, would have to ensure that the purposes of the RMA, and the relevantmatters of national importance, are met through both RMA and non-RMA regulatorycontrols.[98] Finally, we deal briefly with the submission, drawn from Erralyn, that cyclesof river activity should be considered over a period of decades rather than in the short111 Dewhirst Land Co Ltd v Canterbury Regional Council, above n 6, at [43].112 At [43].term.113 Depending upon the nature of the river in question, and the factual issues fordetermination, such an approach may well be justified. However, it does not meanthat the ascertaining of the banks of a river should be of reduced, or even no, relevance.Indeed, in Erralyn the positioning of the riverbed (and hence its banks) was a centralissue for determination.114[99] For all the above reasons we uphold the interpretation of Gendall J inthe High Court as to the meaning of "bed" in relation to a river in s 2(a)(ii) of the RMAdefinition of "bed".The second question of law[100] The Council's position on this question is that the addition of the words"usual or non-flood" into the definition of "bed" represents a significant andunwarranted alteration of the express statutory language. Mr Fowler submitted thisaltered definition cannot consistently be applied across the RMA. It also improperlyconstrains the definition of "bed" and undercuts the underlying RMA purposes.Thus, Gendall J was wrong to imply the additional phrase into the statutory definitionwithout proper consideration of the implications of such an approach.[101] This point can be shortly dealt with. The interpretation of "bed" will requiredetermination of, among other aspects, the fullest flow of the river. In assessing thefullest flow of a river, consideration of a range of geographical and meteorologicalfeatures will be required — see principle (c) at [51] above. What is ordinary or usual,as opposed to what is extraordinary or unusual, for the particular river will need to beassessed. In the same way, what flows are normal or regular for a particular season ortime of the year will be relevant. The abnormal, unseasonal or flood conditions of theenvironment will also be pertinent. So much is a matter of common sense, just as isthe determination of the positions on the dividing line between the banks of the river— principle (h) at [51] above. What is common, ordinary or usual, as well asuncommon, extraordinary or unusual, will also require consideration.113 Canterbury Regional Council v Erralyn Farm Ltd, above n 36, at [28]–[29].114 As the discussion at [30]–[46] demonstrates.[102] In our view there is no need to imply the words "usual or non-flood" into thedefinition of "bed". The contextual application of the definition to the facts of a givencase will involve an assessment of what is usual, ordinary or non-flood, as well as theiropposite circumstances.[103] There is another factor which supports this interpretation. The words "fullestflow" are qualified by the phrase "without overtopping its banks". This can only be areference to flood conditions when the water breaches the banks and flows into theflood plains or surrounding countryside. This suggests that the counterfactual for thiscondition will entail the usual or non-flood, that is ordinary, condition of the river.[104] For the above reasons, we consider that the High Court did err in adding thephrase "usual or non-flood" into the definition of "bed" in s 2 of the RMA byimplication.The third question of law[105] In respect of the third proposed question of law, the Council's position is thatthe assessment of various flow rates or return periods forms an integral part ofdetermining the extent of a riverbed. By concluding this was an irrelevantconsideration, the Council submitted the words fullest flow in the definition of "bed"effectively became redundant.[106] In order to place this question into context we refer briefly to the discussion ofGendall J on the point. The Judge referred to the reasoning of the District Court Judge,who considered the expert evidence put forward by Mr McCracken. Gendall J saidthis:115[55] In his decision, Judge Hassan carefully analysed Mr McCracken'sevidence on the flow rates, including comparing the 50 year and 20 year returnperiods. In the end, the Judge accepted that Mr McCracken's approach in hisevidence was sound and reliable, a factual finding. He then took this evidenceinto account, along with other factors, in applying the legal definition, as hefound it, to the specific circumstances of this case.[56] As I have outlined above, it is my view that Judge Hassan was wrongto do this and his conclusion on the correct legal test was awry.115 Dewhirst Land Co Ltd v Canterbury Regional Council, above n 6.Therefore, I conclude that the Judge did take into account an irrelevantconsideration in making his decision. The evidence from Mr McCracken wasnot relevant evidence that could be properly taken into account[107] We uphold the judgment of Gendall J on this issue, and for the reasons he gave.This necessarily involves a rejection of the reasoning in the second District Courtdecision.116 Any reliance on the methodology of Mr McCracken using the data from50 and 20-year flood returns is flawed. It follows that Gendall J was not in error inrejecting this approach.Result[108] The application for leave to bring a second appeal is granted.[109] The questions of law are answered as follows:First question:Did the High Court err in its assessment of the correct test for determining theextent of the riverbed in applying the definition of "bed" in s 2 ofthe Resource Management Act 1991?Answer:No.Second question:Did the High Court err in adding the phrase "usual or non-flood" into thedefinition of "bed" in s 2 of the Resource Management Act 1991 byimplication?Answer:Yes.116 Canterbury Regional Council v Dewhirst, above n 2.Third question:Did the High Court err in concluding that the assessment of various flow ratesor return periods was an irrelevant consideration in determining the extent ofthe riverbed?Answer:No.[110] The appeal is dismissed. This result follows from the fact that the first questionof law was the central issue on appeal and has been answered in the negative.The second question answered in the affirmative in favour of the appellant involved aperipheral issue.[111] The respondents sought costs. However, this is an appeal in a criminal matter,and we do not consider that an award of costs is appropriate.Solicitors:Wynn Williams, Christchurch for AppellantDuncan Cotterill, Christchurch for Respondents