Ecccgicat APFIICBISO~S Weimesd.>y SPP 05 2C01 03 05 PM 2000 Alien Press . DTPio GALLEY A-223 ecap 12..117 ~p..z23 e

File # 17TQ 00-5140

E<4,#*<<11 ,,,*? ,,,# 2 r. ,I,!,,>>, ,O!j<J,>. ;>p <~l)l~.,i,;,) 0 0000 hi. ihc iiolag:c.it 5,zcrriy at i : n r r i r d

GEOGRt\PIilC T.?\IIGETING OF INCREASES IN NUTRIENT EXPORT DUE TO FU'TCIIE L'RRANIZXTION

J:\\ziis L). W t i - K I I A M , ' * KOIIIKS V. 0'Niii.t.;' K I . R ~ t i . R i : r t ~ i : ~ s , ~ EI.I%ANI:II! K. S t t t r ~ t . ~ T ! t G Wnl>tr ' A N ; > t i K K I I I . ~ J ( ~ \ ~ E s '

iih.srnzir. i r h a n i ~ : i i i o i ~ ii:place< t l ~ c cxt:iilt n:itural icsoilice IYISC? ( c g . , for~.kt:, i iet- lands) with :in irifiastri~ct~iie that is c:~pahlc of supporting huinaiis. One c i i~ l i~g ica l con- iequciwc i l l ixsb:rii~~atii'n is Iiialilii ionceiitretitiris oE nfiregcn (U) ;>rid ph~isptiorons ( t ' i i n s1rc:lriis. 1:iki.s. 3llll ~ S ~ I I ~ J I C : . When ieccivud iii eacei,. N nod P arc coiisidcidd pt~llui:rnti. C~oiiiiniiin:: i i r b a i ~ i r m u n irill cha r~ge ihc relative d i \ t r~h i i i~on of e t a l i i natural iesoxirces. Ch:ir:~cti.ristii.i c ~ f the I~iirlsi-;ipr: can sh:ipe its resprmcc to distnibani-cs such ;is urb:ii~i,:st~i,n. Chaogss in iaiidicapc chsi:ictcristics :rciiiss a region crcare n gcoqiaphic pattciii (if vill- ncrrthiiity t l incic:iicd N trilil P as 3 icsoit ( i f 11rh:inii:rliiin. 1Ve linker1 niltiient-export risk :and iirbimiz:311on models in order 10 id~,ntify :~rc:is fnost v t ~ l ~ ~ ~ r ~ h l e t t ~ i~,crc:~scs ill r ~ ~ ~ t t i c f ~ t - export risk diic to pii,lcctcd urhani~atiori .

: risk-hased 11it~Jc1 of N and t' r x p o i t cxcecding iprcificd thrcstiolds w a s de ie l i~ped h:lird un tlic exl:iiit distrihutiiin o f foicst. :~griculti~ri.. and ilibsn land cover An anrpiiical model of orb;,nii:it~iin mcts used to iiicrc:ise uib:iii land cover at the Lxpenie of ioicst anti apriciiiture The ~nodclcil (Sutti~c) land i o v c r was input into the N and P cxport risk inoiiel. and the "hcforc" :ind "a f t e r ' e5lirnat~!s o f N arid P expoit weic compared to i r l e i i t~ f~ the aretls tiiost vi~lnei:rhIe to cliangc. lncizasc in N and P export had to hc <:quai to or gieatcr th:in the ;iccunlulated uocerta~ntics ~ i l the notiicnt-export risk and urhaniiation mudcis f<>i :In area to he ciiirsiilurcd vultierahle.

I'lic areas mi3st vuloerablc to incic:lscd N arid P export were nor spatially coincidciit with the :mas i,S gieaicst tiib:~i~iz:ition Volrrer:rbil~ty ;also depended oil the geiigrzpliic Jistrihution of iorc\t and :igricoltiire. In general. the areas inost vixlni.r~hlo to increaicd N cxp(>rts i b c r t where the modelcd iiib:lniiztion rate was at least 105% ~3nd the amount (if forest was about 6 limes greater than the iirnoniit o f agriculture. For P, the iiiost voIner;hlc areas were wheic the modeli-d irrbanization ratc was at least 20% and the amount of forest was sbnui 1 tiiiies grcater tlian tile amount of agricuitur:. Viilnerahility to iiicte;isi.d S :ind P export was the rcsull of two intcracling spatial patterns, urbanization ;ind the enr:iiit distrihitt~on o f l:md cuver I t could not be predicted Srilnr either alone.

K r y ilr~rdc erlmit oi iL iiiid P, vuiricii,b~iiiy ;o: ,seii,sinph,i- rnorieiiti~: (;I.$, !nri, i iovei ,Irangr,. iiiiii! coirr: jbri?ii Z / tr~rrculirrrr c:!. iiibiin; irmili<'ii~ir < hnmclui oflerrr ~~oi i r i i~ r~n vtziiirr~~t~ritiv: miii- Ai/<:rtrrr ~ 7 ~ 1 6 , s ~ U S A ) , rr~~~cl~~lt8t~!, ~ r n p z r ~ r ~ t l tmd r t \k tx?.~e<I, t2i:rG:sen; p>zo$p1?ort~,s; re,,qt,,r:~~/~~.c<;i,, ~ 1 r r o I ~ ,v%!s: r1.sk ~ ~ . ~ ~ ~ ~ . s ~ n ? ~ ' r t l ; ; , r / ~ g ~ r ? 8 : ~ l ~ r ~ ~ r t , tw>!trzg cmuf pr<mi#,.,c!i~

lx~rt*or>~:c.~r ION

I:rh:~niiation is the process i>fclranging thc resource base of a 1o;:iIit.; so that is i~i i tahle for higher d r n i i t i occupancy hy iiuinnris. iVhen an nsca is uib:intzed, i t is typii-ally the ilarural rl,soiircc has< that is replaced by ruails, hoincs, 2nd in:titutioti:xl :ind conrmercial strocturcs, Iiciice the resin "de~eiopmeiit ." Fiiri.ts, wetlands. and oilier natural resiiiirces-----:IS well as I"- , e

%lrniis;iq>i r tce i icd 22 2 1 : ~ 1000. re r ised :O Nr>icioiei 2000: ;!cccprcd 29 i)eccrnlici LIXN: final versicii i cce i rcd i Pi.6- wary 2001

" E~m:lil a i i i 1 i ; x i i i i:1:ncsC~~i~g~:aii:31i ep:i g,,i

ricultur;il land--arc replaced by an infissiiuctore cn- pshie of supporting huinaiis, reducing the o~ci:ti! amount of these resources fur :in? z ivcn are:,.

One ecological consequcncz of urhaoiiation is in -

creaseii ciinicntraiions of ilitrrigen ( X i aiid phosphorus (P) irr streams, lakes, snil u\tu:ii~cs 1Omcinik 1977, Beaulac and Reckhow 1982. Frink 1991). S and I' ate cigriificast coittrihutors to cutrophicatton i l fwatci bod- ies (C;npenter i-I a). 1996i. Cutiupiiic:itioil can icsult in decreased oxysen levcis bccaeie the excess lcvcls of N ;and P lead to iiverpioduction o f~u to tn ,phc , iilricli can ttlie:~ten other aqu:itic life (Ct,iiell 1998) S is the fxu~cip;tl Zrgeol o f ei1triipIiic:ltror 8 8 1 ihc occ:rn, f' is the

I Ecolog~csi Appilcations Wednesday Sep 05 2001 0305 P M 2000 Aiien Pmss . DTPro GALLEY A-224

ecap 12-1 17 Mp-224 File # 17TQ

principal a ren t i n freshwater, arid both arc coniributiirs changes <~cciirred. i r c s s of greatest chniii.c in risk were in e t u a r i c s (Corrcll 1908). However. thc role of N in considered rnost vuliiciahle. ciintrihuting to eutrophication in coastal watcrs has The steps were impleincnted hi, ovrriaying o latliic bccn questioned (Ficcky and Kiih:iirr 1988). o f 25-krrr>cells ( 5 km pcr sidc) i>ii the inid-Atlsniic

Studies of the relationslrip between land covcr and , t a l c uf Pcnns)lvaii!a, Xlaryland, Delnuzre. Virginia, the amount of 3 arrd P in wztcr hodits rho\* tliat high ;and \Veil Virginia. Each of tile nii.tti<1doio.ic:1! st~.ps

pri?por!ions o l f;,rc.-1 arc iiceduil tri maintain low w a i pciiormcd on c:ich ccil indiviiluaily. All cells on :rn,oijnt; o f $ 2nd P (Bcriolac ;ind Kcikir~jw 1982. Frink tlie edge o i the stud? arcs i c 2 1 koi') were sxcli~ilcd I')(?l. tluiisaker ct i i i 1092. ilunsakcr and Levine 1'195. iron, the 3ii:il>sis Jones ct al. 2001). The amount of N and P froiri wa- Oiir land-corer d:it;i were fri?li! the Lllii t l~einiiil ir~n ieisl?eds dorn~oated by sithi-r urban or :igricuitiiial "st: I.md Ch:ir;icteristics (MKL<') coiisi?rtiuni (1.iiveiarid are more 5inrilar to c;rch otlrcr tliai~ they are to water- :ind S l i a l ~ 1996) f<,r the l ive mirl-~\tI:intic it:itcs ( \ ' t i -

sheds donrin;iti.ii hy f<irt:st (Rcckhow i.t al. 1080, Beail- gelnilinn ct 11. 1998) l.:~nti covcr was niappcd ilriilg 1:ic :ind Reckliow 11>81. Iziir,h 1901). ,is uihaniiatioo Lands:rt i ?hematic klLlapl?er i'TM) a tc l l i l e data col- prircecds into the future, the magiiitilde of increases in lectcd or, :I t:trget year ill' 1092. i h c i,ri.;loai I'LL ros- Y and P depends not uiily on thc :iiriount o f orbani- olutlnri i i i -30-m pixels 10.0') 112) was rct:tliicd in the i:itio~i, but or, what is ieplacctl hy tllc newly o~hari i r r~d land-cuver iix~ps. i h e tlrcni:itii ilctsii u f l h e land-coier

I~cal i t ics . data *:is rcduced froin hniiersoo l.eiel l i to 1,cvel I Sevcial studies either show or rtrirdcl urhaniz;ltiiin us I ,.\niiersoii e l 31. 1076) hec;lilsc nutr ie~i t - i .a{~o~l data xrc

i(,ntagious (Hockstacl 1996, I'urneret al. 199fi. CI:lrke most widely rcportcd foi urhan. agriculti~re. :itid forest et al. 1997. \%nr e l 31 1998). New urban arcas arc lands. fRere are insirfiicient d:!Ia on iiutrlcot export li?r lrlore likcly to occur close to existing ones. These linii- suhcarvgoiie~ \iitliin thew ~ c n e i a l !:ind-cciier ~ i ~ s c s . inps support longci-staiiiling thearics in eciinviiiic gc- All pixels classilicd as low-dcmity rcsidv~it~:~l . Iiigli-

ogiaphy. such as population distance decay (<.'lark density rcsidenti;rl, :!rid c o n ~ m c r c i ~ l / i n i l i ~ i l r i o l / t r ~ i i s - 195 I) and retail gravity (sse Shi ct a 1 1097). N o i ~ e of purt:ition were trcated as urhan. 1-rkeiiisc, row-crcp:inif these stodics show, especi:rlly a( a rcgional sc;,lc. that pasti>ri, classes ibcre iuiiiped into a biiiglil agricu1tur:il s r b ; ~ n i ~ a t i o n prekrentially replaccs one land-covcr clasi, rind 211 upland 2nd lowland forest clxsscs were type over another. coirverted to a single forest cl:iss. Enrcrgent wetland

Studies covering larger spatial scales arc ~rsclul bc- \+as also iilcludcd in the iorcsl class in this c a c bccaiise causc they often reveal spatial patterir and the fi~ctois erncrgcnt wetlands, like forests. fiinction as liltcrs, lnw- that are likely to account lor that pattern (O'Ncill et crinp niitriant loads in su r roondin~ strc:lros (Prcstiin :%I. IL)?I). Spatial paticin-, revealed by regi,irral-~calc anii liedford 1988). studies can he used ti, inform potential nianagenieiit ii[xioas (Wicktiam et 31. 1999). Our rnairr objective is

~Virrri i . , i i -e . tgi~w ri.rk modri

lo ,how that the rcgion:rl sparial pattern of increases Y and I' riere c x p i e r c d ;IS export cocfficiilnts iii Illis iii N and P is dependen1 ri i i the two interacting spatial study. E p i i r t coefficients csrimate thc mass uiiiutrients patterns o f urbarriiation iiierlain on the existing dis- carried off a watershed by its streams per unit area pel- tiihiition of land covcr The greatcst incresses in ;< and i i r t i t tinrc. Export coeifjcicnts arc thc pioiluct of con-- P are more likely to occrii where urbanjratioii i e p l a c e ~ cciitiarir>n tinles disch:irgc using appropriate :idjust- friiest iatlier tlian i igi ic~~lture . B y undeit;rking the study incnt factixrs. An export coeificient represents the in- at a rtgional scale, it is possiblc to identify localities noence o i 2 given land-cover cl:m when the ~v:itershed

that am irioie vulnerable to N and P increases. Such is exclusively (or nearly so) that c l ~ s s . Fstirn:ites are inii.rmation may be useful for management. often expressed as kilograms per hectare per year

( k g h a ~ ' y i r t ) . and are widely used in the Ii1cr:ituie 'i.lriiions (Oiiicinik 1'177, Heualac and I<eckhow 1982, R;ist ;and

OIIT examiiiation o f the geography of increases i n $ L.cc 1983, Frink I W I . loidan et a 1 1007).

and P due to the intcractiilg spatial patterns o f urban- The amount of N and P exported hln! 3 iv~tershed

i ~ a t i o n and extant farid cover ivss uridertakeir ar 3 four- ihows diitlnct differenczs as a function of a u3ter-

stcp process. First, a riutiient-export risk model mas shed's domin:lnt land cover (fi?rest. agriculture. urhan)

iiiiplernented to estimate tlie pirrbabiiiiy o f N and P (Omcrnik 1977, Heaul~ic and Rcckhi.'.~ 1982, Frink

load5 exceeding a specilied threshold. Second, ;in or- 1991i. Tlic concepliini hasis for using notilent-export hanization model was implcmciited to estirnirte a iutuie cocificients stratified by 1;ind-cover class :IS a basis for

distribution o f l m ~ d c o i c r across a region. Third. the risk w;ls iicrnonstr~ltcd by liartigan et irl. ( 1953) 'rile

nutrient-export risk model froin step 1 was applied to autliors conipared nutrient-erp<,rt coefliiicnts from tlie rriudcled tfutuie) I:irid cover lo vtjriiate the chanec sevurnl iiatcrsheds that were ejther 100'7~ f:oresl, ag- in the nutrient-i:xFori risk :IS a rasult o f iirhnnirsiion iiciilturc, or urban uses uiidcr dry and u e [ years. For Ftiurth, the "helore" ant1 "after" notilent-r:nport risk t'<iicit. t l ~ e dry 1 0 wet year iocre;!re l o r P was 0.04 nl;rps wcrc ct,iilp:wccl ti' ili.teirn,ocd \$here ihc grc; l i~>t kg I?:, ' y r ' ii) 07 In 0.1 I h$-I1:l ' v r ' l . F i , r cropl:llld

Ecolcgiral App:,cations Wednesday Si-p 05 2001 03 05 PEA ~ 0 0 0 I Alien Press . DrPio GALLEY A-225

ecap 12-117 Mp-225 File # 1TTQ

mu nil^ "(HW) I;ORIC:\SIING N ,-\NO i' EXPORT-RISK C'tl:AN(;I<

~

~ ~ ~ ~~~ ~ ~ . . Nutrierti i r c l t p i i i u i r u i e , i i r ? ~ a n . . ( i g h a ' .. - , .~ .

y r - ' , Clin. L l i x Clin. hlnn Mix? Max Sttidy ---- N I 37 7 .?? 1 10 53.20 I SO 38 il l Reckiicou et ,ti L'J%li:

0 0 5 60 5 60 73.40 5 60 38 00 F::~r>k 1'191

. . . . , .~ . , , .

: L):rt:i u c d in ih i i ;!iiiIr .tii.i icr,i,iii.il i n Kccliir,iw ut nl 114F(,) 5 .-XI1 ilala it:piiiicd by Frini, (1 ' i ' ) I i , urcl!idiiig the ilsr;l fio;i~ K c i k h i > u . ei :$I. IO'i) anrl itic

('hc.sapc:%hc l iay

tb;~t restricted tilling of the soil, i l i r i i lrv to wet i i i c r ~ i i c ?L131! 10 /:tiid i o v c r (Rwkirow ct al. IOSO! i t l e d:it:i for P was 0.77 h g t 1 : r ' y r - ' (0.87 to 1 7 1 i\;zha ' y i ' i , lriirn Reckhow et a 1 il9ROi were uscd bcc:ii~sa they hut for cropland that did not rcutrii:l tilling of tlie soil, \\ere i c r e c ~ ~ e r i for consistency and rcpiesci,t uaroril?eJs the dry to wct increase was 12.62 kgh;t ' y i ' (5 02 to ivith homogcnciius (or rieariy s o i l:md cove7 Thcr:ingc 12.62 kg.fi:i 3 . y r I ) . of valucs we used from R L ~ C ~ I I O N el ill. (1080) ax siixi.

M;iny pliysical ;tnd anthropogcnic f:ictors that can iiar lo v:lliles iron1 other studies ('l:rhlc I) . Icxi to elevatiid liutrieiit expiiit tend to hccorne inore The nutrient-expvrt rick inodcl lit the i ~ h i c r \ e d N inkpurt:int as lurest is replaccil hy u r b a n arid agricultuic and P data ( Ihhle 2 i (Rcckhnw ct a 1 I9'iOi to ;I ihc- uses Yet, thc rnagniriide o f tlie impact of tiiese fzictors oreticli ilisrrihution (lognornralj ('lahic 1) . arid iibeil : l i d ii~ierrictioos ainorrg them arc laigclv il i~kniiwn, es- rei~catcd trial simul:irii!n to deveiiip ail estiinate 01' risk pecially at larger spatial 'i'hi~s. i t is approyrizte k*:iscd o n the land covcr in caclr ccli of tile latrice (Wick-

lo cstinmi: tlieir impact a s piohahiiit; (risk) ~ i i m t e i l ham and Wadi: 2000, Wickil:im ct :ll. 2Oi)Oh) A n cijii:i-

ct ;,I. 1 0 9 2 ) Exarnples of some (if t h e e fnctl,is incliide: tion f ror i~ Recktiiiw et al. (1980: Eq. 5 ) b i a s iised to

( l i ye;rr-tii-ye:~i c l l a i i~es i n prc-cipitatjnn (L.ucry :in4 estiinate the risk of cnceciiinp :i specilied threit!iild.

(5oiiIsby 1903). ( 2 ) soil type, 13) sli,pe and slol)i iiioi- .7

phulagy iconvcn, conczre) , (4 ) gciilogy :ll!llon :irid S, 1. - 2 c.;i, ( l i Kirchncr 1'175). ( 5 ) cropping piactlces ( e g , fzictors C , z

;inti P in LISLE [Renard el 31. l'J971). (6) timing of li~i. I estinrates N or P cnport a s the prodiict o f tlie icrtiliier applicaticm relalivi. to p~2cipit:irion events area (;I! of land-cover type i tirncs its export coefficient (Beaulac and Recklrow 1982). ( 7 ) density ii i impervi- (c,) suinrned across :ill lanil-cover types in the cell. i t oils surlace (Arnold and Gihbons 1096). (3) resideiilial each itcratioil, a initiieni-export coefticicnt (c,) w:is lot sires, arid ('1) iatio olscwered vs. septic ihokieholds. drawn lor iiaih !and-cover class in r sch cell osing the

Our nutrient export risk model tWickh:lin i.t : r I oicitn and rari;incc es t imxcs in '1.abie ?. N iinri P vc rc 1000?>) was developed from a bi)dy of literatiire tli;it then caIriilatcJ as an aiea-ivsigliicd avcragc using tile rci3te!i N and P eapr:rt ( i n kilograms per t l e ~ t a r c per propoitioil o f cach land-iokcr class in thu cull. B y using

T I . 2 Ch;increiiii~;s of nuriieiit Itinding d+rz iisi.d i s > i 'eiel~ip niirr~col~cx~ioir risk inodcl. ih ta are i iom i<i.clihow ei ;a1 i I ' J bO; 1ioIdi:icc i i i i i i i t > e i i were oscd as i1,icshoIdr 11, derclop risk eslim3les~

-- 0bscivi.d iiaiixrn: export d:,i.

ikg It:% ' ~ y r ';+ Wsteiihcd Xn $,I -

l.;,nd co ic i s i c h i xorrii-ni s I Q,, Q,,, Q L 1 . 3 ~ ~

hlired agriciiltiim .I:!-di)00 N -30 1 10 0 60 I I I i ) 2 0 .30 53 L O i' 17 0 08 0 4 ii 01 1 3.l 5 4 0

R3r i . i~ 7-47 000 N I 1 :%i 1 92 2 411 3 .;Z 7 32 I> 6 O l l t 004 1108 U ? 2 0 5 3

i 'Thc c i ~ l o m n i:c;liliiigs rmgc ircm m,it!miim ihi<;iigl~ inorimum, Qli. Q;, V. . arc kwer c~u:*r~~lc, mcd>:m, ,inti tuppex q ~ ~ . ~ r ~ : l ~ : . rcspcct,\?l>

I Ecotog~ca: Ap~licut~ons Wednesday Ssp 05 2001 03:OS PM 2 0 0 0 Allen press . D T P ~ O GALLEY A-226

ecap 12-1 17 Mp-226 File # l7TQ

8 ; I . Llap of risk of N export exceeding rneiliiin value f o r uibaiz 1:ind c , ,vei iQ, - 6 5 k g h a ' - y i ' ) . A intiice of 2.5- km: cr!!r (5 h m pcr side) has been overlaid o n ihc m i d Arlnnric st:ites oi I'~no~y1v:ini:r. >lary!:,oil, l>clow3ic. Virginia. anJ U'rst Virgxnia Boidcr ccils '25 kmz in arcn are black

I:iod-cotcr percenta$es, the units of the equation are thu slime 1: the cxport coeihcienl itself ( i e . . kilogram. per hectare per year). Randomly draivn cxpart ioe f f i - i icnls (r , ) iicic discarded i f they iverc oijisitlr thc ran-e o f the abscrved data (see 'Fable 2 ) so that tile risk cs- timates wuoid be conservative.

'7'1,:: iandoinirarion u:rs ~;pca!cd 10000 Ilincs for each cell. 'The riuiiiber of tirncs (oil1 01' i 0 0 0 0 ) th:~t !he eiiirnatcd U- i i i P-export coclticieoi equalcd o r cx- ccedcii the incdian coetficieirt value ii,i urban land c o v ~ cr (6 5 for U. I I O for P; see 2) was used as a mcasuie of risk.

There :ire no unerjuivocal choiccs for sciecting 3

thrcshold its a hasis for estimating r i s k Rast ;and Lee 11983) piiiposeil -3 k g h a ' - j r ' for N 2nd a r:ingc o f 0 0 5 ri i 0 . 1 kg-ha ! y i ' for I' :IS n3iion;*lly (Lloited Stntcr) icpiesentativc vniiies lo r ioiests. Flowartii el 21.

( I 'Nb) p ~ o p u ~ ~ I 2.7 k,g ha !!i ' :IS air tipper lirnrl tl'r

N export for Si,re\rs under pro-Europe:%" settlement ciinditioni for the nnithu:istern LTnited Siatcs, hut did not ,xopo<e i tlrreshuld ISor pliosphorus. Our maximum value for i' i x iic:ri the iipjicr end proposed by Ras! and Lee (1983). hut t l i i l proposed values for U iR:ist 2nd Lee 1983. tio~.v:~iili el nl. 1996j are low cornpareil to iinr valncs Oiir i?bsei,cd rncilian valocs i o r tirhan W:I- leisheds ~ p p r o n i i i ~ a l c thc ilpper limit of 3 3011 P 2nport froin iorcited iv;l!er4icds for thc datii \be used, and

rcllcct ti,:: ie:rlit) tlisi 100'7, forest cannot he :xchii.ved unifoitnly 3cn)s i 3 i x g c rq i i ?n .

l,o,r,iicnvt~i iriiiiiitio,, tr*o<fvl

rnixlcl far I:IIICI-C(IVCI change w3s es!ahliihed throiigli an eriipineal icl:~liiinship hctwecn the amoillit

01' iirban :ind road :ihiindniiic in e:ich icll . Tlie pro- partirin urh:in pixels per cell from the / ~ ~ ~ - C O L C I i131:1 bias j i i ~ t t ~ d 3 m n s t i.::ch c::/i's road i i tns~ ty tK, in

I Ecoiog,ial Applications Wednesday S r p 05 2001 03 35 PM 2000 Allen Press - DTPro GALLEY A-227

ecap 12-117 Mp-227 File # 17TQ

FIG 2. M a p of risk of I'expurt urcccding median value for arban i:iiid cover iQ;, - i i kg ha-' y'). Foim;lf -is in Tiz. I 1

.. .~

meters pcr squnre kilometer), and fit with a mudiiicd al!omctric function (Parton and innls !'J'Z) iusins scg- inented nonlinear regression:

P(Y,,I L') : 0 ( S b i R 1 60)

P!Y,,/ L') = 0.003(R - 00)0'9' ( h r R >. 6 0 ) 11)

Ey. ? estimates the proportion of p n r l i ih:it werc identified as urban ( P ( Y / i/i) 3s i i / j incti~it (olfltc c e l l s road density (R) . The prtiportian of urban pixcis per cell %as estimated as the actual count diuidcd hy the ~naximurn possible nurnhcr f f ~ i 30-m ( 0 09. ha) pixcls in a 25-km' cell (27 775 piasls). 'Total ro:id leiigth pcr cell was estimated using U S Geological Survey (USGS) l : i00000-scale digital Iinc ,graphs (I>LGl for rranportarion (USGS It1Y9). ,,\ittioiigii ievni:li ancil- lary getigraphic data sets were used lo dcveliip thc l:jnd- Lwver srlaps (Viigclmann i ' t a l ~ 10981. road\ &cre

iiscd iri the cl:tssi!iuation (J. E. Vogelmaiin, ,iersoriui c<>>,,m,',,;c<lli<i")~

Fq. 2 was scgrnentcd into two ciises to :iccoont i o r thc cnisfcncc of iu.ids when the proportioil o i urban lp~xcl; in 3 cell u a s 0 A l l o ~ n c t i i ~ ~qu :~ t i i i i~s , peih&yi most commonly used in popitlation modeling, assuiiie the <lc:pc.riilciit v;rri:iblc is 0 whet\ the independent fac- tors arc (i This :lssurnption was not valid for describing the relatioiish~p be!,.vci-n road density and proportion of iiiban pincls bccause roads ofteri ciinncct polii~latcd places by p:issini: through relatively ui~seftled areas.

The piedicte~i vlilites from Eq. 2 for ciich cell were uscii to d i i i e land-cover c h a n g e For ~ 'naniple , iS thc piedlcrcil value Tor :i cell was 0.2, tileri LO'i o f the pfxcls riot classified as urban rieie changed to iirban I,!rh;:n incre:lrcd nt the expense Sorest and ngriciiiture. i,!rban cuuld nor incrcsse at the expcnse OIU:IICI- Olhcl Ianii-cover caiegorics ( c g , bare rock/s:iiid, inining,

Ecotogicat Applications Wednesday Sep 05 2001 0 3 ~ 0 5 Pirl 2000 I Arten Pgess . DiPio GALLEY A-228 ecap 12-117 Mp-228

File # 17TQ

Eiot;,$!crt Appl:ial:nes JAMES 5. VlICKHAM ET A L ~ V<%t. Cil. Yc, C)

1~0:) , 'r,tiii.~ 3. Mczii and v:tii:$ncc estimates for li,i!nuiiii:$l dii- trihctiuns i>S U- nnd P-zrpiiit coefiicicnis. hy taiiil-cover 0.95 I class.

093 -- .-.- 0.95 Land cover Nutiiciit Mu Sigma

- 053 - Llixcd u:riculluie N 2406 0911 0.75 - P - 0.22 1 I 0.56

:! 07C ~ Urhon N 1 000 ( iOt i P 0 .233 il 981) 0.65

Forcsi U 1 021 3 5 0 6 P -2.351 t Illi

in urban was taken froin torcst :ii,ii 40'1 was takzil from agriculture.

Eq. 2 was used as a model o f t h e predoiiiiii:int pattern of urhaniiation in the castern United St:~tes. Thriiogil- out the 20th century. urbaniiation has fi~llo*ed a spa- tial pattern of deccntralirati<,n-uibsn coorei i i im spreading fiotn an existing urban corc ((iioliano l4l9'1i Road density is used here as a iiirrosiitc S b i this pio- cess. The total length of roads tcnds to be higllcr at

0 100 201) it10 ,100 500 600 700 locations more proximal to urhan centers. 7'he :Iccess provided by an existing, more well-developed road ilet-

Road Density (tnikn~:) work fosters a hisher prohability of future urhanizatiiio. FK;. 3 Proportion of urban vs . road density Pis-

dicred t:ilucs am sbirwn in bhitc unril road density equals Ill0 m'km? in niiter to disringuisii from observed ralucs.

tinnsiiional) were ignored. These categories tended to occupy only a sm;tll proportion of any cell (<0.01). Eq. 2 was restiicled s o that it could only change n

nnximum I 0 000 pixels per cell, about 35% of the cell's total number of pincis. The estirt>atrd change frotn Erl. 2 servcd as the proportion. The amoiint of pixels ilia1 could be changed i l 0 0 0 0 ) scrved ;is a length of t i m c ~ Combined, they provided n rate of change per cell.

The potential 5570 change appears to he cr~nsistcnl with other land-cover ch:ingc studies in the region (hat spanned zhiiut a 70-yr period. 1,;iiid-covcr change es- tiinatcd from temporal change in noriiialized-difference vegetation index ( N D V I ) for USGS hydrologic ac- counting units (watersheds) was as high as 30% hased on imagery acquired for the targct years of 1972 and 1990 (Jones ci at. 1997). Esiiniatcd change (l'rom NL>\;I) averaged across the entire rdgion %was 20% for the same period (Wi~.kham ci a 1 2 0 0 0 ~ 1 . The diftkrencr it , thc two rates for the same region is a functliiii of the reponing unit used. Smailer units will have gre:itcr variance and greater likelihood of high rates o f change in some cells. Our 25-krn2 cells are inuch sinalier than USGS liydrolupic accounting units used by Jones ct ai . (1997). There are about 125 USGS hydrologic 3c-

- . This trend in the spatial pattern of urbanization has heen incorporated in existing, more lucal-sc:rle rrihan- ization models (Clarke ct al. 1997).

Esrirnarin,~ of vulnemhiliiy lo N ond P irtcrclies

Output froin thc transition model yielded 1;md-cover distributions with higher amounts o f urh;in and iowcr arnoririts of forest and agriculture in cr l ls whcri' the modeled I-ate of land-covcr change was 1 0 . Thc inod- eled land-cover distributions were then used 2s input into the niitrienr-enprirt risk model. As in the fii.sl 3p- plication. thc nutrient-export iisk modcl was run tlirougl~ 10000 iterations pcr cell, and the tiumber of times the estiiiiatcd N- iir P-export coefficient ciloaIt4 or crccciled tliz riredinn value for iirhan w:lo used as a tncxsurv of risk. Vulnerability to increases bit N :%rid P was ri~ensuied 3s the difference in iisk ostirnates be- tween 'hefc>re" and "aftcl-" map,.

"Bcfor-" and "after" nutrient-export iisk cstimatcs were calibmtcd for random effects in the nutrient-cx- port risk iiiodcl i n d cnor in the proportion urhai, vr

rooti ilcnrity regression equation. The marinium dif- feicncu in nutrient-export risk for cells wlrose i>ri.dictcd orb;lnizalion was 0 were 0.026 for ?4 and 0.019 for P

Tni i i .~ a . Gooiintss-of-!it cstinvic for pii:poition <:I' uil,an pixels v s road iicnsily.

Sum , i f >lean ~ - ~~

iounting units hut over I1 800 25-km2 cells i i i the re- soiiicc d l sqiiaies squ:iic - gion. K s g r “ w n (16.75 18 3 8

Tht- total incrmse in urban ~ i x e i s in ;I rcll reu1tt.d n,.,,,i,,,i $ 1 8 5 8 15 17 0.00117 in a proportioi~nl decrease in tiwest and agrieulturc. I f Toi:,! 1 i 800 I I I .92

the proportion o f Sorcsr pixels in a ccll w : ~ 60'i o t the I I "' 9').IO -- silin of forest and :igriciilturz, then 60'Z of the incic:~sc ; ('~,rrcciioo h a s ~ ~ ! on rhr nican vS1hc iictlcniient viiiahlc

tcolagicill A p ~ t ~ ~ ~ l i o n s / Wednesday Sep 05 2001 03 05 P M 2000 Alien Press . DTPio GALLEY A-229 ecap 12-1 17 Mp-229

File # 17TQ

POKECASI'ING N A S 0 P EXPORT-RISK CH,\NiiE

The it;urdaid error for tire individual picdicted valiiec valites thzrn arcas diiinrnated by 3giicnlioral use br- in thc urban vs. road density rrgrcssion equation w a s c a m e the observed i!:it:i SII( IC(CS~ that N-cnpi:rt cocifi- 0.0358 Added togetlier. these unccr i~int ies total 0 0 6 1 cisnts tcnd to he higher uttcn :ireas ;ire dominated h y f o r N and 0.055 io r P Diffciences in "bei,re" :tnd agricultt~re tlian ib1ii.n ttiey are dominated by urban a f l c r ' . ~i:k est imulcs h;,j to be c q u a ~ to or thair (Bcaii1.i~ iind Reckhoiu 1982. Prink I99 I : set! 'lhhles 1 these vaiuci iu hi. consideicd vuinciablc. and 1). 'Fhe IL'wc:~ estciiiates or risk :ire wliere fcicsts

donrinntc. including north-cenii:il Pcnnsylvsnia and Rrsui, i c snuthein West Virginia.

E~tiii,,i niiiiizr,i o/.V /in,/ P rriiiri~ni-rmijri risk The sp;brisl p:ittcin iifi ' r isk [Fig. ? )d i ik r s somewhat ~. ~ .~ ~ ~~~ ~ ~

X I t f ' N 'i'ilc cstirn:itcd i i sk is higher for urhan Thc risk nf N-export coefficiziits eqoniing o r ex- .,,, lha,, for :isrii,,llur31 arc3), hecause the ,,bsCrve,j

cceding the median value for tirhan ranged from 7 C r 0 data susgest thilt ~ . ~ ~ ~ ~ ~ t C(ic f~iCicT,ts are sliyhtly lijgh- (Oi to about 0.77 (77%) (Pit . I!. The areas of higllcst ei ovt.r;,~~ for urban :ire;ii ,\isi>, tile risk estimates are risk are ihe Ccntial and Shcnandoah Vtlleys of I'enn- lo,,, th:ln the r i sk c s t ims t r .~ iiii 1. because P-export sylvanis and Virginla (respectively). ttie Piedrnonl of coefficients for foi-cst l i o w leis overlap with either soullicnsiern Pennsylvania and central Lllrrylz~nd. :~nd urhan or agiicultoie ihil N-e.\port coefiicients. Esti- the l)elm3ivn Penii~sula. I'tic high vnlucs o i risk in m:tted f'-export C O C I ~ I C I C I I ~ I :ire I I IUCI I more sensitive to tliese :ireas rellect tlie \sidespread coi~veriioi? of 1:riid tlie :iniount o i iorcst i and :rmount o f forest removed) to :figrii-i~ltoi:~l u s e i:ih;ln :ircas tend iv hzvc lv.icr risk tlian X-export c : ,cI l i i ient~~

I Ccohgical Avpticaiions Wednesda, Sep 05 2001 03 05 PM 2000 At!e?n Priiss . DiP io GALLEY A-230

ecap 12-1 17 Mp-230 File # 17TQ

J:t%IES D. WlCKtij\L1 El' AL.. licotog~cai hppiicarsrri VOI. m m.. o

Fii; 5 \i ruincrabiliiy overlaid on iiihoniz,isun map iFtg Ji. Oiitlincd cells identify :iic:ri whcie S risk ~neicased h i at le3it 0 062 3s o irrult of urh;tnii>tion

Lorid-cover rmnsiriorz i\ MSt\ is an area defined hy the 1:s. Census Bureau 3s lmving at least 100000 people. Nearly one-fourth Tire ielatioiiship hetween the proportion 01' pixels of the region's 25 &1S.\s are concenrratcd in south- classiiied as iiihaii and road density \%;is si$t>ifican! eastern Pcniis:~lvania, which is only about 5% of thc ( F i p 3. '1:ibIc 4) TIic proportiori uf urhair pixel.; in a tiital Iaisd arca of thc five states. cell reached 50Ci. at ;i road density o f ahoiit 320 in:

knr' Thc spati:xl pattcin of the propurlion of urban pineis appc:ired to accurstcly ,nap the m;~jor cities in i t i o n F i g . 4). In addition to sfiowitig t i l t i:irscr population ceiitcrs (Pliii:~delphia, B3ltinioic. LVasliii,$- ton, Soriolk. Richiiiond. Pittsburgh). i e \ e ia l ~maiic i - urh:in centers >.vem also evident ( e g . Koani.ke ;rtid Clilirlottcsville, Virginia: ;\ltoona, Harrisburg, ;and York. Pcnnsylv:ini:1: Bcckley :ind Moigantown, West Virgliiiz). Cells with proportion of urban pixr:is -:O5 hut :.0 I tcndcd to surioiind urban center5 I.hc con- ceniraiion of cells *ith tire proportion i i i urb:rn ptncls -0 10 in .;oiitlic:~stcin Pennsylvania rellccts the higli iiiiniher 01' t1~~tropolii:in Si;rtist~c:tl Areas (MS:\ ) thcie

For N, 43 cells had :in increase iri risk equal to or cicatcr than the :zccutnulatud i>nccrt:>inty in the modcis (0.062). 'These cells are not spatially coincident with the areas o i greairst proportion o f urban pixels p fig^ 5 ) . ,Clost of thc areas where the proportion of urhan pixels was :>0.50 Rnd ch;liisili in N risk less than the 0 .062 threshold. Many o i tiie areas where N r3i.k in^

creased hy 21 least 0 0 6 2 werc where the proportion t r i

uihan pixels was lower. 'The main exccpiion to this \pati31 p:jiiein \&:is i n tile vicinitv n i Pitisburgh Ll~ich

Ecological Applications Wodne-,duy Sep 05 2001 0305 F'PJ 2000 Alien Press DTPio GALLEY A-231

Month 00('0 FOKI~CAS'IING N A N D P EXPOKT-RISK CtIAUGE

ecap 12-117 Mp-231 File # 17TQ

fii; 6 1' voiociability ovrilaid on urhan,?oti i i i i map t i ' ig . 4). Oulitncd ril~.s identify arcus where I' r isk incrc:~rci! h y :%I lens1 0 !J55 a; a rewit of i i ibsnizai~on.

of tire area still lias a high proportion o f tiirest covcr i V o ~ ~ 1 1 n a n n et 31. IOUSI. As ifrb:inii:riion cvntinucs i n

tilib area i t is likely to i,ccoi at the cxpeiisc of forest. For f'. 385 cells had an inireasc in risk equal to or

~ I C . I I C I Illan the accumulated ~incertairlly in ilie models ( 0 0 5 5 ) . 'riiese cells .how greater spatial i>vcrlap wit11 those wilere the proportion of ~irhnn pincls uas grc;itest (:zO 5 ) (Fig. 6 ) Nc~eitticless. riisny ,if tiic cells wilere f' risk increased by at least 0055 were in :lic:is where ihc l:kc!ihood o i urhani~a!ioil was less l'hcr: wcic a greater numbi.i of cells u u i i i c r ~ h k to a change in P than U because the oomeiic3l slrparatiim c>ienport co- cffic~enrs between forest and cither agriculture o r urban was more distinct for P tlian N.

l 'hc spatial interaction of uib~ni ia t iui l and relaiivc :ibund:ince of different typcr o f land cr,vcri created a geographic paltern of N and P vuinci;zb~lity tirat i-au1J niil be pwdicted riiiri, eittrci iniidcicd iiih;~nirativo or

cntant land cover s l i>nc Tlie 1andi~:ipe ua'; irioie viri-

oeial?lc to incicased P c-tporr ii.iien the tiiodeled dis-. tiirhancc inh:i:ii/atluo) was :about 0.20 or higher arid tile 1:~nilicope hsd ahon1 ? times inore foreit tlian ag- ricnlturc (Fig. 7j. For N. the laiidsca1?c was rirorc vul- nerable when the modeled distuibnncc =;IS shout 0.20 or highcr and !he 1andsc:lpc tiad :rhoiit 6 l i r n r inore fi,re>t tlim :iericoltuic (Fig. S ) For 3. decrc;ises in risk eiiual ti, iir sreatci th3ii t l l i ai.ct~n?i~latc(i ilrrcsrl:~intics in the rni:Jcl i 0 . 0 6 1 ) were :liso iiiiind ulii'n the ratio of ioiest to ;igiicuiture was very low a i d rnodeled ur- h:rriiiation \\as !iigh ~ Y C C Fig. g) ' S l i ~ : tvailahl~ liter- ature su,ggcsts that a~ri~uI111rc p~cscn t s tiighcr o ~ c r 3 l l risks ,>ie\i.ess,vr N exports than iiitnn aie:is (Beaulac and Reckhow IOK?. Fiink 19'11). 'l.liz iesporrse lo ur- hanii:xrii~ii by landsc:~pes characti.iizi.d hy iri~lcll gre:ilcr ami,unts i>f : )gi t i~l t i i re might be Inii-r i,vei:lll N c x p o r l ~ hI:iny ;~uiicaltui:il sicas me ~ i i s c ~ ~ x ~ b l ~ t i t i~rb:tniint~iii~

Ecolng~cal Applicaiions 'rYednesdey Ssp 05 2001 03:C5 PM 2000 ! Able" Press . D.rPio GALLEY A-232 ecap 12-1 17 Mp-232 i

J

File # 17TQ f F

F,,; i ~ e ~ : , l , o n z h i p ~ r i w e s n ci;aii..c ~ , . c r p ~ i i i r a t i o %;I' forest LO agiwb~lriiii- v i . pii.poriion uih;mircd Solid c,ii)ils idcnt,ii mo-l v,,~nei:ih#c 1; ihengi. ,,, p ( ~ h m ~ r in risk incienved i t 1cs;t 0 (155). Plus rigns tilentif? i.rt(s .ihi;ie ;h:tn.c i n risk was ,cO 055

( l~lcaiy arid Short I'JSL). 'riris directii,n in rcsponse ii:is not apparent for 1'~

D t s c u s s ~ u u

Expressing the landscape as an n-diiiiensional set o f charsclciistics ( s . ~ . , Figs. 7 :ind 8) has heen useful for cxarnining disturb:trrce icspnnsc (O'Neiil el al. l980. 1904. Turner ct al. l'ii13L The landscape i s likcly to

respcrnd diffc:irnily as :I function 01' the extent oS the dis torb:z i~e i e l i ~ ~ r v c to ihe cricnt oi iiie 1nndsc:ipe

('Turner ct a 1 1 0 9 3 ) Figs. 7 :lnd 8 show that ihc land- scape response was a lunction o f the size 01' the dis- turhunce (increased ciihan l311d) relative to the atriounl of forest. As increases in urhan land replaced greater arnounls o f : ~ ~ i i c u l t u r e . the landscape was less vulner- able to increased N and P export.

The geogi:iphic paticins of increiised N and P cnport risk shiiiv a dichoiuniy ivhcn overlaid on 3 illap of lire ~ . a t e i s i i cds o f the innjni rivers i n the reaioil (Fig. 91. ,\bout 70"; o f the cclls ideutified as \ulnciahle iu in-

I i t l i t r r ~ i ~ i , t i l i i l l "5 0.15 0.25 0.35 0 4 5 0 5 5 0.65 0 -5 0.85 (195

Proportion ijrbaitired

1 6 . Kelni ioohsy I7ctwcm i-h:,npc in S-export i l s k ani! rzsiiv ,,i iuicsr ti? agriculture i s pn,puitiun iirb:,ni,cd Solid circlls iiien!iii cel ls rnt tc t vi!lni.izble t i , ch;%ngc in N e ipor i ichzngc in risk incrc:%,ed 31 least 1) 0 6 2 ) Soid iii:inglei idrntiiy cells whose change i n n s i . ~!circu.cd a t !cast i) Oh2 l ' liic signs idcii!aiy ccl!; whose change !n iik ;vsr beticen -11 062 ;and (3 0 6 2 ~

crcnscd P export i icci~iizd in watersheds that diz' . iii 10

thc 4tlantic. I'hc piedorriinaiit pattern of increased risk of P export follows tlie urban corridor from t'hiladel- phi, to Richmond. For N. the majority o fce l l s showing vulnerabtlity t i , ~ n c r r i s e d export drain iiito thc Ohio K l v e r 2nd iiltirnatclv suotli to tlie Ciulf of hlexi io

'Tile actual amount of increased N and P load rhst iiottld re:ich rhc \ t iant ic or Gulf Coasts is dependcnr i!n the s i i s i l l tlrc >iri:3ir>s iihcre nutrient expurr occiiis

(Behrendt 1'2'16, Smith et 21. 1 9 9 . I l i c in:rss o f nu- lrients retiloved fninr streairis ttirougil dcniriilication. sedimentation, 2nd ottiei processes i s inveiscly rclated to streain size (Behrendt 199li. Smith e! a 1 i99 i ) . Tiie plocesscs lh31 rcrnove nutrients fiiiiir rticains tend lo be more effcctivc in sinailer streants. ,Alexander el a!. (3000) eslimalcd that 20 to 4 0 5 oi the N exported from uatcrslrrds in the mid-AtI:inric di;i in~ng lo the Ohio River ir:lched ilie Gull i,S 'ilcnico.

I E C O ~ O Q ~ ~ ~ I ~pp l i~n t ions kwednerday snp 05 2001 03 05 P M 2000 niien Pies$ DTPio GALLEY A-234

ecap 12-11? Mp-234 File # 17TQ

our resu l t s , like a i l modelin. stiidie,. 3 r ~ . d t p c n d ~ r l t on tile :lssuinpt,ons built into i!,c nii,dei.. Cliai1:ing c?ne or ,he assumptvons wotild likely change the

,rhe ihieshoIds useii !ii Jctine vulncrabiiity for

N P wcrc cirnicriative, 1 .e . ~c lec ted LD ~ C ~ U C C lhi. of ;l false p<ositi\e. i'or e\aniple. the :iccLI-

m,,lalcd u n c c n a i n t y U S C ~ I [he maximiin1 d~ i fe rence froin ti,e n u l r j c , , t . c x p o i ~ risk m ~ d c l when ilic c h a ~ i r e in the

ou,,t of o (oo:~. x: 0 0 1 9 . P!. e s i v 3

icss (,,,\ ei,,at,vi. ,~:,iue (c :. 05th peici:nii1ci ~"o i i ld

lr:,.c incresi-d the niiinber o f cells identi!icd 3s vilj- nerahlc.. thicshold of 0.05 inbicad o f O 00'- 1'01 5 wt,uid !he number o f cells ideniificd 3s

,,lnerablc. ~ i k e ~ i ~ e , changing ilie IhrcshoIdS that

UsCd to risk (bold nuinhers in 'Pdhle 2 ) ii,,uiii chaiige the results. In add~tiiin. chanqll%! the

to convcri forest and spriuuililre 10 ~ ~ r h a "

c o u l J changed ihc rel:~tivc P x ' p ~ f l ~ ~ l ~ f ~ C , K S ~

,ind niriicu)tuie ccinveiteil tt, iirhan. :inil I).<IICLI ~ l ) a l j ~ ~ d

I Eculcgical Appl'ratiors Wediiesd?, Sep 0 5 2007 03 02 PP.f 2000 Ailen P i c i s . DfF7~o GALLEY A-235

i lcnth tH?If) I:OREC'ASTiKG 'i AUI:) I' EXPORT-RISK CfIhNGE

O n e c c o l q i c ; l f i o n s e q u e n c e of i i r ban i~a t i on is in-

/ , . . crzascii cxpa r t i?!' N 3nd P to re,:elving w atcrz {Oiiierilik

197:, Hs:iulac :iiiii I<cckftorv 1982. l'iink 1991). \Vc developed a moilel l u eslirnate I ~ P risk of N mid I' expor t ciiualiiiii or cxcceding specified thresholds hnscd o n rhu ert : lni distribot,iiii o f - land cuvcr. l i i e e r r zn t larid

e , .zs . i,r.iiigcd .i u s i n s 3 model to io i iue i t lei-st and : i g r i c u l t i ~ i ~ ti, i iihan. and tile icsullairi l311d i i i vc r W:LS

a.;tJ iicii.ce iicxr eiiiin:itus of N :$rid P risk ;IS 3 rcsiilt

o f c i>nre i t ing non-i irhsn land ti) o ihan use. ' 1 ' 1 1 ~ J ~ f - fsreiicu in t he risk eiririratcs was used t o idt!ntify ttic

ore::s inosi u i i l n e i ~ h l c r i i iiicreased U :urd I' cxpur i 3s :I resrilt i i f 3 Siirurc urhnniz:iti<iii o:ittcii~. Areas \+ei., considered miis! vu i i ie rah i t to incrc:iscd N ;and P export wilci, If~c risk ts1itn;ites cfiangcd hy 6.2'2 and 5.5',7, r c s p i ~ ~ ~ v c l y . 7 'bcie t t i icsi iol i l wc i e iiciiiivd by rhc :ic- C U r ~ ~ ~ I a t c t l L:IIC:~I~:~IIIIII:-, in llic i i i i t i i t i i t - t xpo~ t riiitl ,IT-

b~i i i i in r loo moi le fs ,

' l 'he g t r ,g iaphic pattcrir o i aie;ls iiiosi wuincrable to inc~c:ised N- :ioil P - e rpo i r risk was the result of. the

iliteractinp i p ; ~ i ~ s l pi l i t i ins of ttrh:lniiarion and cxi:int I n c T h e iiiocr v i~ l i i e i i~h l c a reas occiriieri ibhcie

i h r piojecred uihanii:iti<iii r:ite *:is 31 least 20',::. :,nd the riniciiint of Siiicii %:IS about 6 tiiiias greater fiian

itle nmvi~i i t of : ~ ~ r i c o i i t ~ i c fiti N, :ind li,ri.st iv3s 7 tirilcs

grcntcr lhan ~ ~ I I C ~ I I I I I I C f 0 r P t :rhani~:~ti i , rr wil l likely c o n ~ i i i o e lo i p r ead , mo\t

pr~i lo i i i ioznt lv in :I core-to-pcrlphery spati;ii pattern

i I 'The izsionaf partcril o f 1:md-c.i,vci

ch:lilgc is like!? t o h r s t rongly inlliicnccd h i ilie coic to-pc i ip i i t iy p:litein of i ~ r h a n ~ r a i i o n (Wickham e t at .

20000) Witt:ir, iiiis pot<tiitial p:rttern of oihaiiiiatii,ii

arid 1;lnd-cnber clraiige, those a reas most ru1ner:ible to

ecap 12-117 Mp-235 File # 17TQ

13tat~l:tc, .X %.. ;ant1 K f l l t c c k l ~ ~ ~ ~ u lW2 . An ex~n?tn:tlion of land ? ~ C - I I I I I I I C I I I i x / i i i i i r~l?liitilsll:p~ W:itcr l l r in i i rc~s IiiiIleTiii 181013-1024

h i if. I . lnvenroiies of po~n t arid diffuse iourcrs :ind csiiiiiarrd notiicnt I ~ m d i - - - ~ c i i m p i i s u n ic i different iibci I~iwiils i n criirial Eiiiopi.. W l e r Science ;ind Tech- nology 33:99-i07

Ilocksluci. N. i. 1996. Mcdsling rcoi;omici and cciilogy: the imyoilnncc of ;i itprtizl peispicri,.e. :\rncrbcan J o u i n ~ l nf ?~gr~cii l luial E;onoii,i;s 7 8 1 166.1 1:O.

i'nrpcnrer, S K., N I: ('.ii:,cu, i ) I. ('oiii.11. R. i\' liw.vari!~, T?. N Sh:irplcy, and V 11 Smilli I O ' J X Ni.npcxnl p,~llurii~n of surface watcii wiii> ;hi~ijiiir:riis ;ii:d rtilrogcn Ec01ogic;il h r ~ p i i i a t ~ ~ ~ n s 8 559-568

I . 1'451~ t.$rb:n p o p t ~ l t > ~ ~ ~ > n dcns~ties ! ~ ~ ~ ~ r ~ : t l of the 1Coj.ll Siaiisricai Scciciy, Sciies ;\ 115140-496

I . ti S I I I J I 7 A self- riiixlifying crlliiloi .liiroiir:,ti,n moilel of hisriiiicnl iirhanr- Iariiin i n the S z n f : i a n c i i , , B a y A"". Envw'.iimo,t and i'l:inning i 3 PI:~oni;~g and Dci5i.n 24247-261

' r . I I. I'i98. rhi: role u i pilo-phorilr i n the ~ t i t i o - phicnlii,o ,if iecesvmii h:aieii. a i c i i cu , louioal ,,I E i ~ r i - ii~nmental Ooal!ty 27 361--266

t)illon, C I . a n d W 13 Ksrchoei 1'175 Thc uffcctiofgu,,logy rnil laad kist on the export ~~fpt iory l~o ' .us from w:~teishsds. Water Kcsearch 9 - i 3 i - l 4S .

Fixnk. i'. K. I Esrimatinf ntitiicnt r tpoirs to cstii-iiics~ Jc,uinal of Enviii,iirnenraf Qimlic? 20717-72.1.

( ~ i t i l i ; l i i ~ ~ , G 19%9 Lztiid w e poiicy and tianspl?it:it,iir~: why we Jon', get rheru lii,m here Transpci1:~rion Kcsearch C i i ~ celai 495 179.198

iiailigln, I I:. 'C F Quascnh:liih, iird E Soi~thurlsnd 19% ('aiihrarion of Ni'S moilel loadiiig facrors Jiwrnal o f En- viribnmciitul Engine~i ing 109,1259-1273.

licnly, K. Ci . and I I. Shi>ii. 1')SI. 'f.ilu inarkel for rum1 l ~ n i l trends, iisuei. j r , l i c i e~ ?'iti. ( '~~oic iva t ion Fiiunds- Ison, Washangion, I) <:. ('ST\.

Hccky. K I;, :!nd F'. Kilhnm 1988 i i i i r i s n t lirnitzrion oi ptijioplanklon in firsliwstcr :and z?i:tiinc eniironments ;I

i t s ' i e u cii recent ci!dencc tin the elfccis of cniichmect i.~iiinulogy :,nit Occinogiaphv .43-79f>--X21.

Rciendse, i Frcney. V Kudey:%iov, R %liiidi>c, ;inti Z Z11::o- A< KXOWI.I . !X; \~I 5 , ~ s L.iang 1091, Repion:ti nitrogen budgets and riverinc N &

Tlir icscurih il-pi>itud i n this p;iper u ; t i fi!ndcd by ihe I! S Eniii~~nincnI.rf i ' iiiic~iioi~ \;ii:ircy li liac liicn siil>;eci I:, reci ic>ic-, atid ~~!mizi~.tiai?vcly :~rie;vcil 1.y the ; \ ~ e n i y S:ippiii! f i ~ K O 'Xc~ l l \was lhr<ntgl? ~nreragency ~ t g r e c t ~ ~ ~ ~ ~ t UWi'i9;Xii-01 112 ,w,Ih Oak liidge Sniioi!:rt i.:~ber:iiiiiy \ o y mcniiilii i,i iiailr. ii:>mcs docs noi coiiiri1ii:r cndcirscoicnt i;i iicirrni-rei,daiii,n for t i c Tlic iiirlic~ii ;$re gratciiil to >Ion- 1 i 3 T i i r ; i ~ r arid ti-.- .IIIIIIIYIIIOIIS ~CII I I IYC~S for their C O ~ ~ ~ C I I I C

tin r:niiiv: vir.;i<rii i rhe n;ti.cr

~.ITI.R,?T,:IZE Crr t z

I i . R 13. R I Stnsth, :and (; E Schw.>r, :O(AJ i i i c2 i ibi stirarn ctiaiiiic.1 rice on the dcli icry rrf nttio.cn lu the (ilrlf of Lisr:cri Nature $0.3 758-;h!

?ndciitiii, i I: . l E t!aidy. J. T ito:iilr. ;,ad K E \L~liiici iq).:r, ..\ land I : ,~CI c c ) b < r ~ I ~ , s s , T ~ c : ~ ~ ~ ~ , , I sys:,:rt, :<,r ,lie w i t h remote sensor d:it:k 1.: S (>coliigicnl S u i i e y I'io- icssi~~n:il P:;pcr ':h4 tI S i;oi;iniiieiir Priiitine 0 1 6 ~ ~ . \\:*sl>,,l~!<>,,, I> (' , l ;s>\ .

I d . C L.. i ' J I t I impciuiiius ,i~if.r;;

1/1c i i i i c i ~ c n i c < s t .b i c y ciiviionmeiiral iiidic:iti)r li>iiriial of riie . \ m c r ~ . ~ > ~ ~ ~ ~ : , n , , ~ ~ g A < ~ , > C ~ ; : ~ ; ~ > , , 62 :~.a-2i:

li:~ilcli, S XI.. i i I 1 (;drdnei, anit K \' (!'Neil1 1 Ec:, I<V?!C,I I r>sk es1tx,,313,>,, LC",*>, <'i,~l,c;,. >l!~.I,,g;,r,, l":\

P t 1 i i . r ~ ~ 10s the d:;,inngci i,f ihe 51>1111 :\tl;tntii. Occon nniiiinl and hiirn~:n i n l l i ~ c n ~ r . liii,gci>cheoiistiy 3575--1 3')

Iliiiiuairi, C T . :tiid D \ 1 , r i ~ i i ~ 19'15 kticizlicli~czl p - primilhcr lo rriidying na re r ,jii;iliiy iii iivcis. i3ioS;icnce 45 193-203

iitiiirnkci C T. 1) 3 l.c~inc. S, P Tii:imini, i3 I , J;;cksirxi, aoii i t V O'Ncill I . L.anJsc:,pe cli:~i;,ctciirn!iiln fur ;issciiiiiu ie:ixoniI ,.vatei ijiia1:ty Pa.ies 991-1006 i n 1) it &IcK;.niie, 1) E. Flyair, :and J J >l:rclhioold. ed i tc i s Ecii~ logical indicali~is. Elcciiei ipi:Iicd Science. Vc;v Y<,il;. Nc,w Ycork, i,'SA

i:~ccs. K 13 . : \ C' ,Yc:$le, 1 1 S hash, K I> Vast ltc:nririrl. J I) Wickhtrin. K It Kiiticis, ;iiJ I< L O'Uexll 2i:OI Pirdici iog iiiilricnl a c i icdiincat Iu:tilingi ,a itrc:~-ns from lantlscapc incirici- a rn~i1t:plc :uirei$hed study tioiii !lie United S!:ites. L3nJs;apc E;oli,gp ih:?Ol-312

Joiics, K 13 . K H Rilrieis, I U iVickbam, K . I). T~nlersle?, K V ONe; i i . i ) I Chaloiid. E. K Smith. n d :? C N c l i c ~ 1 7 I\ ,) ~ c i i l o g i c ~ l assessiiicnt id the I:iiiisd St:iiei in id^ AcI:>nt~c rcgton. (L l ~ n d s c ~ ~ p ~ ;<tias EI~.\;(~OO/R.'17:~30 OfBcc i i i l<cscaicli .md 1XvrIopmcnl. I ' S Eiiviii~iirncnt:l 1'rori.crioii i g e n c y . Wai'iingliin, I! (' . llS,4

Joidnn. r il . I> 1. ('iiiiell. and I) E \Vcllii. i ' ) ' ) i i t fcc ls t i f ~ g i i i ~ i l l i i i r o n d,ich;rige$ i d o i l i i i c i i i s fi<m? C<,;sii:il i'1::ir

I Ecological AF~IIC.II,OI?S Wednesday S e p 05 2001 03 05 P M 2000 Allen Press . DTPro GALLEY A-236

ecap 12-1 17 Mp-236 File # 17TQ

w ~ t e i s h c d s ;tf (hi. (:kesnpi.she 13:~:~ Jcii:n:,i of Eiiri ioo- nicnrai Quality ZhSY-dl&

Li?vclaiid, i' T< . a n t i 1) Sl S h a ~ v i'i'l(i. Llulii-icsniiiii,iii !.rn(i ch. - I I ~ L I C ~ ~ ? ~ ~ ! O T I , . -hua!iiiiip ~ ~ l l n h ~ i a l i ~ e partnerillips 1':ig- es 75-X5 t r 3 J ?+I. Scott. 'T ti Tear, and F \V. U:IV$S. ed$to'< Gr\P analisis-a 1andsc:~pc ippii>.ich iii biodiversity pian- "in.. .Amr i , c ;~ :~ Si>srci: of Phoioe i~rnrne i ry and Kcrniiic Sens$og, ilc!l$csd;~, SIarylamJ. l:S..\~

1 , K J 2nd I : ; I 1 . Eliecis ~ i f c l i r n a i r c v;~ii:~tioiis o i i i I I yi:tis on niiiaic-niiiogce con~ei i t ia i i i i i is i n the Kacili,ri K i b d i , 10iva. Soiiiilid ~i E11~11~nln-.1:ill ()113/-

Lccti,," , \ ~ C , > C > . <',,r~;lIl,$. O , ~ $ < , , I , IZS , \ . O'Nezll, I< i. \ K Johnrr:n. 2nd ,\ W King i8)Y9 A

hiei-archic:tl fr:tini.iroik for the ::nalysls of sc:iic L.~nclica~ic Ec~i lugy 3 i 1 i 3 - 2 0 5

O 'Xr l l i . I< V . K ii Jiinci, K F( Itiitiris, J 1) iVick1xirn. and I \ (;;ioiiiii:rn. I L:%nd$clpe monitoring .ind :is- siscinrn! rdsc:rrc!i plan. i1'Aib10!11~91/011L) Oillcc t11i i i i : ~ su:irch ;and rjcvcl<rjiincnt. 11 5 Env#ionii~eni:,l i'ioiccrii~n A o c i ~ c r . W;isiiingrvn, i? f: . 1151

O ' U C ~ I I , K i. S. s ' ~ u i i ~ i . ~ , v I f'iliiinan. I) i' t 'atiin. 'r Cook, \V Coii!ci, J Iliun!. I b! Tiiornas. $1 R Cor!Icy. a d S i 1901 \liiiliplc 1andhc;ipi. rc:iiei aii iirtcriiic cn;np:lrison i.oniisc:ipc l:coiog: 5 I?--14.4

P3iioii, 1V I , and <i S lnnic 1972 Somt. ;ii?phc :+nil i l i i i i inoction;il forms 'Tccl~nical Repiiii 153. li S. liiivi;iariui~:~l !lii,logii.al I'iopram, i j in \%lasd Riorni., C'iilt~rsdoSI:?ri. l ' t i i >cr$iiy, Ftmt C t > l l ~ n > , ( , ' ~ ~ l t ~ r ~ ~ i ~ ~ . l!S:l.

I'iesron. E. h! . coil i f . I. Bcilioid. $OR8 E*olo:itiog it$-

mitlaiivc cifecii oii xcr1;iiid irinciions: I uonceprrr;ll nrc i - view and i.cnci>c fraiivcwisrk Enuimnrncnt>l Llan,gcincnr

lakei Jouin;iI of i n v i i o n m ~ n t a l linginccriny. 10% 5 0 1 - 5 I7 lli.ciho;u, K 11 . 41 *; I3c:itilac. m d J T S ,mps iv 1')AO

L!~~ilciiiig phoipirtiiiis luail$ng and lake icspi.nic iin4i.i u n ~ ccitaiiity 3 "1311131 anti i.;,mpil:ttion o i rxjioii coc:ticiciiti l \ i l S - S l i t I I ' S E;iuimomeiiial l'miectioi? ,.\i.iii- cy , Wnsh,iig!on. 1) <' . llS.\.

Reii;iiil, K ( i . <i K 1:iicter < i ,\ uleesic., TI. K. hlc('<x,l, cod 1.1 C Yodci l'i')?. gtiidc to caorr:v:iiion pii~iiniiig

~ .~ . t i l> ilw r c v i x d unb\ers:%l SOP! loss c:jtixtw~ (Ri 'Sl-E) .Ag. iicii:iuid tiafidhoiiii So 703 il 5 !)cg;aiiiiieni <i!' ,\gr*- ctiliuic. Wohinpion. lit' . U S A

I . Y I 7'. . ' h s a I I I h : i i iuI lui~l land \oliici under uib:sri;,io. inliiienccr Land Eroiiornicr 7s 90-1 0 0

S t I \ , (i E. Schws i i , and K R A l c x ~ i i ~ ~ r ~ I W 7 Keg$on3I ~nieqirc<;itioo of v.alerqiz:~!i&y mtm,ri:iiii:: riai:r~ Wr ic i K~.soilrics Kcsearch 13 27ti1-2;98

'Turner, L1 I [ ; . W t i Kornii!~. K t i (i:~idiici. I < \ fJ'Ncili. ; , r ! ~ ~ ~ ~ i i 4 i ~ i i i i , t i I . ~ ' ~ ~ ~ ~ . ~ I I:U)~ISC;LDC. cil!nl~hiiiiiii. disiiiibance :tiid sla!itlir) on ic:llsd l:iniiic:ipc3 I..xni!uc:iirc E c a l o ~ y 8 2 13 -211

7'iirnci 11 ( i . 11 N i V e a i :,"<I K 0 Flamni l'I'l(> l.:ind o~.*$>ii.ish,p a n d land-.-o*ci ih; inec i n ihe i~ii:rIlnrn Appla- ch i an 1~i:~Iil;inds :ind !lie Olympic penili>iila I<c~ilr,gicaiAp- p l ~ c ~ ~ ~ ~ ~ ~ n s 6 I I?O-.l I?:

I)S(;S (if S ~ ; c o l i l p i c ~ ~ l S n i i w j 1 1 i>i~ii:il l ,nc g i a p l ~ r < r t m I ~ l i ~ O O l l l l - ~ c : ~ l c rn;>t>s. l>:$f;t l . s e t ~ (;u!dc 2 1,; S~ (;vt,-

plic;~r;oiis 3 blii--GiO \Vtckh:sn?. J 11, K 13 Jones, K t i . IZ~!!h;rs, 'I' f i Wade, : ~ n d

I< i O'Ycili I 'lli:,nsiii~~ns in < ~ , r c i r lirugii,cntaiion. i i i l l l l i ia t i i~n: ; fix icirr~r::rion uppt,itiiiiiiic~ :lr irgii,a.il sc;ilr:s. I.;,ncisc;i,>c Ecology 1 4 1 3 7 - - 1 1 5 ~

\V,ckh:irn. I 1) . I< V 0 I K I . O < J . A ,ueo2r;8p~ly ui ca3sy~cc;~r vt~Ineralvla!y l . ; ~ ~ ~ ~ l s c : ~ ~ ~ ~ Ecctl~>$? s .

1.122 ivickharn, S I , I T I V : 2000 Sii;tti.?l iiatlein of

irntcr p01I:it~i.n i i i i i r i Llarylsnd. I lSA Pages ~~~~~~167 in

Pii,cecilings iii the 2nd In1einarii.n:~l Cc,nleiciiie on Gco- ;parial lnfi~irnarion i n Fi3iciiiy and Agriciilliiie, Lakc f ioe- na Visi;i Fli.rrdn, I:Sh. Janu:iry OOO ERlLI lnicinotii~n;tl, , \nn ;\ib<ri Slichignn, l :S - \~