North American Journal of Fisheries Management 16:257-271. 1996 American Fisheries Society 1996

Historical Trends in Abundance of American Shad and River Herrin Albemarln gi e Sound, North Carolina JOSEP . HIGHTOWEE H R National Biological Service. North Carolina Cooperative FishWildlifeand Research Unit 7617,Box North Carolina State University. Raleigh. North Carolina 27695-7617.USA ANTON M. WICKER AND KEITH M. ENDRES' U.S. FishWildlifeand Service Post Office Box 33726. Raleigh. North Carolina 27636-3726. USA

Abstract.—The Albemarle Sound basin once supported large fisheries for American shad Alosa sapidissima and river herring (alewife Alosa pseudoharengus and blueback herring Alosa aestivalis combined); however, current landing f botso h stock extremele sar y lowusee W .d calch-per-unit- effort (CPUE) data from an 1845-1907 haul-seine fishery to estimate the historical abundance and productivity of these stocks. These historical data are valuable because we can measure (1) the respons stocke th f eo s toperioa fishino n f ) stoco dg(2 durinkd productivitan Civie r gth Wa l y before significant changes in water quality and habitat. We found that biomass-based models fitted the historical haul-seine CPUE data reasonably well t thebu , models appeare unreliable b o t d r efo forecasting. For river herring, an analysis that included 1977-1993 pound-net CPUE data appeared provido t e better management advice. Estimate populatioe th f so n growth rat ) rangee(r d fro5 m0. to 0.9 for American shad and 0.3 to 0.5 for river herring. We estimated that maximum sustainable yield was roughly 1-2 million kilograms for American shad, compared to 5-6 million kilograms for river herring. If the stocks can be restored and current productivity is similar to historical levels, both stocks should support much higher landings than are currently being made.

American shad Alosa sapidissima rived an r her- Carolina 1980). Our goal was to compare estimates ring (alewife Alosa pseudoharengus and blueback of sustainable yields durin e 1800gth s with current herring Alosa aestivalis combined) once supported landing asseso t s potentiae sth l productio f thesno e large fisheries in Albemarle Sound, but landings stocks. have declined substantially from historical levels (Figur Chestnu; 1 e David an t s 1975 . WinslowS ; , Fishery Development and Catch Analysis North Carolina Division of Marine Fisheries, per- Development Historicalthe of Fishery sonal communication). For example, 1994 land- American shad and river herring were among ings of river herring (253,000 kg; Winslow, per- firse th exploitete b fis o ht d commerciall Nortn yi h sonal communication) were only 5% of the aver- Carolina because their oily flesh allowed them to age 1880-1970 landings (5.4 million kilograms; sale b t preserved withou r refrigerationo e ic t e Th . Chestnut and Davis 1975). Landings of American establishen fishera s wa y d industr time th f eo y yb shad have declined almost continuously from a Americae th n Revolution. Between 177 1776d 1an , peak of 3.3 million kg in 1897. The current low the Edenton, North Carolina, custom house cleared landings have led to concern about the status of 851 vessels carrying 24,432 barrel fisf so h destined these stocks adequatt ,bu e recen availt t no dat e -aar Britise foth r h West Indies, the Azores Canare ,th y abl evaluato et factore eth s responsibl- de e th r fo e Islands, southern Europe, and the New England cline (Winslow 1990). colonies (Taylor 1992). The objective of this study was to estimate the By 1846, there were 15 large haul-seine oper- historical abundanc d productivitan e f o thesy e ation Albemarln si e Sound employing about 1,000 stocks e analyzeW . d catch-per-unit-effort (CPUE) people (Taylor 1992). These operations marketed data from the Greenfield fishery, which was a haul- their most valuable product, American shad, in seine fishery that operated wit same hth e gead an r Baltimore, Maryland, whereas river herring were procedures from 184 190o 5t 7 (Universit f Nortyo h shipped to markets in Richmond, Petersburg, and Norfolk, Virginia (Taylor 1992). 1 Present address: U.S. Environmental Protection Agen- New markets were established afte e comth r - cy, E-MAP Mail Drop 75, Catawba Building, Research e Albemarlpletioth f o n d Chesapeakan e e Canal Triangle Park, North Carolina 27711, USA. from Currituc Norfolko kt , Virginia 1869n i , e Th . 257 258 HIGHTOWKR ET AL.

4X1O6- American shad

3x1O6- I |2x106

1x106

0x10°- 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990

1x107- Rrver herring

8x1O6-

6x1 06-

4x1 06-

2x106-

1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 Year

FiGURt: I.—Total reported landings of American shad and river herring for Albemarlc Sound from 1880 to 1993.

canal enabled steamboat transportatio Norfolkno t , the passag e sounth r df migratinfo e o g fish (Taylor Baltimore, Philadelphia, and New York (Taylor 1992). Apparently, before passag f thio e s law, 1992). Ice also became available during this pe- spawning runs were completely blocke panely db s riod, allowing fresh American shad to be trans- f nettingo . ported to northern markets at premium prices Haul-seine operations rapidly disappeared after (Boyce 1917) responsn I . developine th o et g fresh the introductio f pounno d nets because haul seines fish market, the fishery was expanded with the were very expensive to operate and could be fished introduction of pound nets in 1870 (Earll 1887). on only a few sites that were free of snags. In The number of pound nets increased from 117 in comparison, pound nets were relatively inexpen- n 189i 0 0 18895 (Nort o t 0 h Carolina Boarf do sive to operate and could be fished anywhere. Agriculture 1896). By 1896, there were 1,125 Many were near the ocean inlet, whereas the haul pound nets in Albemarle Sound and its tributaries, seines were primaril western yi n Albemarle Sound resultin estimaten a n gi d tota f 242,21o l f o 0m nea mouthe Roanoke rth th f sChowao d ean n rivers deenettinm p3 7. (Nort go t fro 7 hm 3. Carolin a (Figure 2; McDonald 1887). Also, the haul seine Board of Agriculture 1896). Use of pound nets and industry had focused on salt processing of Amer- stake nets proliferated to such an extent that, by ican shad, which became obsolete wit availe hth - 1905, North Carolina found it necessary to establishmen e pasth d s an abilite ic f dail f o t yo y rail- the Vann Law, which require channeda l through road and steamboat connections to northern mar- HISTORICAL TREND ALOSAN SI ABUNDANCE 259

08.11.

Blackwater River

FIGURE 2.—Locatio e Greentielth f no d haul-seine fishery, which operate n Albemarldo e Sound from 184o 5t 1907 historicad an , lspawnine rangth r efo g migratio f Americano n shad.

kets. Fifteen haul-seine operation- s Al wer n o e afted e r propert wated yan r rights were purchased bemarle Sound in 1846 (Taylor 1992), 12 operated in 1882 (State of North Carolina, Halifax County in 1880 (Boyce 1917), 4 were functional in 1896 Superior Court 1907). The dam was on the site (Stevenson 1899) y 190b ; 2 onle Greenfielth y d occupiew Roanoke thano th s i t y db e Rapidm sDa fishery remained, and it closed by 1907. (Figure 2) and averaged 3.6 m high, creating an Another factor havy thama te affected landings 18-ha reservoir (U.S. Army Corps of Engineers of America lose n spawninf th ssha o s dwa g habitat. 1943). McDonald (1884b) believed that loss in McDonald (1884a) reported that migration up the American shad spawning habitat would result in Roanoke Rive blockes rwa t Weldonda , North Car- correspondina g reductio productivn ni e capacity. olina, whereas the natural limit to migration was An ye fisher th los o t sy attributabl a los f o o st e furthem k ove 0 r r25 upstrea m near Salem, Virginia spawning habitat would not have been evident in e (FigurobstructioTh . e2) t Weldona havy ma ne the 9 fisheryears3- r ,fo y becaus enume thath s -i t bee nfisa h slide, whic s describehwa Yarroy db w ber of years it takes for adult American shad to (1874) as a solid, substantial structure that cap- return from the ocean to spawn (Winslow 1990). ture l Americadal n shad attemptin migrato gt e fur- Direct effects of the fish-slide harvest at Weldon ther upstream. Later e Roanokth , e Rives wa r e unknownar t Yarrobu , w (1874) suggested that damme Roanoke th y db e Navigation, Water Power the harvessubstantiao s s wa t l that American shad and Manufacturing Company e specifiTh . c dats ei would disappear from those waters withiw fe a n unknown, but the dam would have been construct- years if that method of fishing was continued.

TABLF. 1.—Number of American shad fry stocked in Albemarle Sound and its tributaries from 1877 to 1881 (North Carolina Departmen f Agriculturo t e 1881) recordo N . available sar r 1882-1884efo finae stockine th ,year3 lth f so g program (Winslow 1990).

Numbe f Americao r n sha stockey dfr : din Location 1877 1878 1879 1880 1881 Roanoke River 389,000 350.000 150,000 Mehcrrin River 150,000 230,(XX) Nottoway River 111.000 275,000 Blackwater River 220.000 550,000 675.000 Chowan River 200.000 Salmon Creek 1,508.000 210,000 1.075,000 1,860.000 Albemarle Sound 600,000 185.000 110.000 875.000 1.570,000 All 600.000 2,543,000 1.395,000 2,650.000 4.105.000 260 HIGHTOWE ALT RE .

1845 1850 1855 1860 1865 1870 1875 1880 1885 1890 1895 1900 1905 600-, American shad

1845 1850 1855 1860 1865 1870 1875 1880 1885 1890 1895 1900 1905

River herring

1845 1850 1855 1860 1865 1870 1875 1880 1885 1890 1895 1900 1905 Year FIGURE 3.—Haul-seine effort and CPUE for American shad and river herring in the Greentield h'shery. 1845-1907.

Concern abou effece th t f fishino t American go n tributed to the increase in landings in the late 1800s shas firsdwa t expresse mid-1800e th n di s (Taylor (Winslow 1990). 1951) e approacOn . r addressinhfo g that concern prograa s wa stockinr mfo g American sha inty dfr o Description Greenfieldthe of Haul Seine Albemarle Sound (Winslow 1990). The stocking Operation program was implemented from 1877 to 1884 by The Greenfield fishery was 22.5 km west of Ed- the stat f Norteo h Carolin U.Se th .d Fisaan h Com- enton alon e nortgth h shor f Albemarleo e Sound mission. Numbers of fry stocked in Albemarle (Figur ; Stevenso2 e n 1899; Universit f Norto y h Sound Roanoke th , tributaries it f o e e Riveron s r o , Carolina 1980). It was owned by Edward Wood, ranged from 600,00 187n 0i 4.105,00o 7t 188n 0i 1 who also owned and operated 2,025 ha of farm- (North Carolina Department of Agriculture 1881; land, served as president of the Albemarle Sound Tabl . Althougt possibl1) eno s i t o separati ht e e Navigation Company that owned steamboatd an s the effects of stocking from the effects of the ex- schooners, and invested in the Chesapeake Canal panded pound-net fishery, stocking may have con- Company and the Seaboard and Roanoke Railroad HISTORICAL TKHNDS IN ALOSA ABUNDANCE 261

3,5x106 -

3.0x106-

6 B 2.5x10 -

12.0x106-

| 1.5X106-

H 1.0x106-

5.0x105-

0.0x10° 1845 1850 1855 1860 1865 1870 1875 1880 1885 1890 1895 1900 1905 Year

I seine

0.0x10° 1845 1850 1855 1860 1865 1870 1875 1880 1885 1890 1895 1900 1905 Year FIGURE 4.—Total reported landings (upper panel) and estimated landings (lower panel) of American shad in Albemarle Sound, 1845-1907.

Company. After Wood's death in 1872, the Green- annual catch in number by the number of hauls and field n fisherFrank so s operate s wa a yhi , y b d multiplyin estimatn a y gb f meaeo n - weighin r pe t prominent businessman who maintained interests dividual (American shad: Stevenson 1899; river her- in farming, banking, and manufacturing. Thus, the ring: Smith 1907) (Figure 3). Greenfield fishery records presented in this paper The fish were caught with a 2,285 m X 7.3-m are those of a large, integrated industry that was haul seine with 3,29 f haulino 0m g line (University well maintained from 184 o 1907t 5 . Alsoe th , of North Carolina 1980). The seine was laid out Greenfield records were made before income tax by two steam-powered flatboats starting from a was instituted, so they would not reflect any effort fixed location offshore and hauled in by two steam- to avoid taxation. powered engines on shore. The fishery employed Microfilm copies of the original fisheries ledgers, about 80 people to work the seine and process the made available by the Southern Historical Collection catch. Seines were operated for 6 days and nights of the University of North Carolina at Chapel Hill, per week during the run, which was generally from were examined to determine the number of fishing April 1 to May 10. Typically, four to six pulls of hauls (Figur annuad an ) e3 l catc numben hi f fishro . the haul seine were made each working daye Th . We calculated CPUE (kg/haul) by dividing the total number of American shad and river herring were 262 HIGHTOWER ET AL.

7x1 06- 6x1 06- IsxIO6-

I?4x106- •o « 3x1 O6- |5 2x1 O6-

1x106- 0x10°- i i i i i i f i 1 § ii

1x10

Non-haul seine

Year FIGURE 5.—Total reported landings (upper panel) and estimated landings (lower panel) of river herring in Al- bemarle Sound, 1845-1907.

recorded for each pull. Records of American shad ,., = rBB+ t t- (\B f/K)- C, , (I) landing t includno d di es hickory shad Alosa me- diocris, whic e similahar appearancen i r , because Bt represents biomass at time t. r is the intrinsic hickory shad were considered inferior in taste and rat f populatioeo n growth, s unfisheKi d equilib- wer t marketeno e d (McDonald I884b). Alewife rium stoc kcatce si/e th Cd t tims ha i , an , e(Hilr - and blueback herring records were not kept sep- borWaiterd nan s 1992). arately because the two species were (and still are) We modified this standard model to account for sold collectively under the name river herring. On the effect of the developing pound net fishery on the basis of analyses for several Atlantic coast riv- haul-seine CPUEe differencth n I . e equatior fo n ers, alewives and blueback herring have very sim- biomass, we included separately the catch from ilar estimated growt d naturahan l mortality rates, haul seines (C//, d othean ) r gears (Ct0y primarily maximum sizes, ages at first spawning, mean fe- pound nets, but including staked gill nets): cundities, and population growth rates (Crecco and BHI = Bt + rBt(\ - B,/K) - Clh - Clo. (2) Gibson 1990), so a combined analysis may not be a significant source of error. Because most pound nets operated downstream of the Greenfield haul-seine operation, haul-seine Analysis of the CPUE Data CPUE (Us assumet tima t )wa e propor b / e o dt - We analyzed the historical CPUE data using a tiona biomaso t l s remaining after harves othey b t r biomass-based model (Hilbor Walterd nan s 1992). gears was accounted for, e standarTh d model assumes logistic population U, = a(B - C,<,); (3) growth (Schaefer 1954) and can be written as a t difference equation, the parameter q was a proportionality factor rep- HISTORICAL TRENDS IN ALOSA ABUNDANCE 263

1.4x107n Haul seine

1.2x107- I I Domestic other gears

B| Foreign fishery landings 1.0x107

0.0x10° 1845 1855 1865 1875 1885 1895 1905 1915 1925 1935 1945 1955 1965 1975 1985 Year FIGURE 6.—Estimated total Albemarle Sound river herring landings by gear category. 1845-1993.

resenting haul-seine catchability solviny B . g equa- (Hilborn and Walters 1992). We also obtained ad- substitutind r Ban tfo tio ) n(3 equation gi n (2)e w , ditional least-squares estimates for q, K, and 0\ could expres difference sth e equatio r biomasnfo s by fixing r at a range of values. This approach, in term f CPUEso : sometimes termed profiling, illustrate welw ho dl the paramete s determinewa e datr rd th an a y b d the objective function, and how sensitive the other parameters were to changes in r (Mittertreiner and Schnute 1985; Prager 1994). The biomass-based models required estimates of annual landing geay b s r category. Record f totao s l

Clm0). (4) annual landings were available occasionally between 188 d 195 d annuall0an 0an y thereafter (Figure 1). In additio e logistith o nt c model parameter , K,r s We used linear interpolation to estimate total annual and q, we also required an estimate of 0\. the landings for those years when data were unavailable. predicted CPUE leve yean i l r one. Approacher sfo We estimated landings from haul seinin r 1845gfo - obtaining 0\ include estimating it as a fourth pa- 190 multiplyiny 7b g Greenfield landing numben i s r rameter or setting 0\ equal to U\ or an average of the first few (/s (Hilborn and Walters 1992). Al- by mean weigh r individua pe ty linearlb d - an in yl though estimating an additional parameter is dif- terpolated estimates of the number of haul seines ficult when parameter confounding is high (Hil- operating assumee W . d that non-haul-seine landings bor Walterd nan s 1992) obtainee w , d better fity b s were negligible before 1880 (Milner 1880; Earll estimating all four parameters. We used nonlinear 1887) and equal to the difference between estimated least squares to estimate the four model parameters total landings and haul-seine landings between 1880 y minimizinb g and 1907 (Figures 4, 5). We tested the reliability of the 1845-1907 model 2 ,- u t) (5) for river herring by predicting abundance and 264 HIGHTOWBR KT AL.

-•- CPUE

600^ ...... Predicted CPUE (1845-1874) - - Predicted CPUE (1845-1907) 500-

s-400-

§300- Ui

O 200-

100-

1845 1850 1855 1860 1865 1870 1875 1880 1885 1890 1895 1900 1905 Year

2.5-

1.5-J

1-

1 0.5-

1845 1850 1855 1860 1865 1870 1875 1880 1885 1890 1895 1900 1905 Year FIGURE 7.—Observed and predicted CPUE for American shad in Albemarle Sound. 1845-1907 (upper panel) and predicted biomass from the 1845-1874 model (lower panel).

CPUE after 1907. Because model predictions were mated with the standard formula for a logistic poor (see Results), we also fitted a model to the model, (Kr)/4 (Hilbor Walterd nan s 1992)- ad n I . entire 1845-1993 time series, using 1977-1993 dition to the estimates of MSY, we have also in- pound-net CPUE data (Winslow, personal com- clude r estimatedou f historicaso l biomass. Those munication n additionaa s a ) l inde f relativo x e estimates must be viewed with considerable cau- abundance. A separate pound-net catchability (

0.3-

0.25-

& °'2~ J 1 0.15- LU <° 0.1-

0.05-

0.2 0.4 0.6 0.8 1.2 r 10000-

K (millions) 1000- MSY (millions)

100-

10-

...... ^...... 1- w H •

0.1 3 0.2 0.4 0.6 0.8 1 1.2 r FIGURE 8.—Population growth rate (r) versus error sum of squares (SSE; upper panel) and versus estimates of K and maximum sustainablc yield (MSY; lower panel) for American shad, based on the 1845-1874 model. resulting from fishing t occurrebu , d primarily dur- were obtained for r values of 0.5-0.9 (Figure 8). For ing two periods: 1874-1876 and 1901-1907. r = 0.5-0.9, estimaterangeY MS d o f t froo s 9 m0. There appeared to be a moderate increase in CPUE 1.8 million kilograms (Figur. e8) after the Civil War (1861-1865). Haul seining was prohibited r becaus durinwa e gth e North Carolina River Herring authorities feared that the foodstuffs would be ac- Similar to CPU r AmericaEfo n shad, CPUr Efo quired by Federal factions (Taylor 1992). river herring varied considerably from year to year e biomass-baseTh d t provid mode- no ac d n di ela (Figure 3). Some of the fluctuations in CPUE cor- e entirceptablth o et t CPUefi E time series (Figure responded to those for American shad, which sug- 7). A much better fit was obtained for the period gested thacause th t e coul similae db r fluctuations between 184 d 1875an 4 (Figure modeTh r . fo le 7) in year-class strength or changes in flow or water that reduced period adequately capture e fluctudth - clarity that affected catchabilit f boto y h (Figure ations in CPUE before the Civil War and predicted 3). There was a sharp increase in CPUE for river a sharp increase in CPUE after the Civil War. For herring after the temporary closure during the Civ- the range of r values examined, the error sum of il War, then a decline to a level similar to that of squares was least for r = 0.7, but very similar results the pre-war period (Figur. e3) 266 HIGHTOWE. AL T RE

Predicted CPUE CPUE

1845 1850 1855 1860 1865 1870 1875 1880 1885 1890 1895 1900 1905 Year

25n

15-

10-

5-

1845 1850 1855 1860 1865 1870 1875 1880 1885 1890 1895 1900 1905 Year FICHIRH 9.—Observe d predictedan d CPUE (upper panel predicted an ) d biomass (lower panel r rivefo ) r herring in Albemarle Sound, 1845-1907.

The biomass-based model provided a good fit late 1940s. No data are available to evaluate the to the 1845-1907 CPUE data for river herring reliabilit e 1908-197th f o y 6 CPUE predictions. e mode(FigurTh . l e9) predicte dshara p increase The model predicted a sharp decline in CPUE after in CPUE after the Civil War. then gradually de- hige th h landing e 1970sth n i s , which agreed well clining CPUE afterwards obtaines bese wa t Th .fi t d wit available hth e CPUE data. Parameter estimates = 0.5fo r whicr r fo ,estimate th h f MSeo s Ywa for Ihis model were considerably different from 11.3 million kilograms (Figure 10)modee Th .d ha l those for the 1845-1907 model. The best fit was poor predictive power, however s CPUa , s Ewa achieved at an r of 0.3, for which the estimate of forecast to decline to 0 by 1911 (Figure 11)7 millio5. . Theres nwa kilogram-Y MS s (Figure 12). fore, we refitted the model using 1845-1993 land- ings data 10 accoun continuee th r fo t d existencf eo Discussion the fishery. stocy ke ke assessmenOn t questio whas ni t sort Predicted e 1845-199CPUth r Efo 3 modes wa l of analysis is warranted, given the available data. similar to that for the 1845-1907 model for the For American shad and river herring, we consider period up to 1907 (Figure II). After 1907, the the haul seine CPUE data to be of high quality. 1845-1993 model predicted substantial fluctua- The annual values represent a large number of tions in CPUE, including a sharp increase in the hauls with consistent gea methodsd ran valuee .Th s HISTORICAL TRENDS IN ALOSA ABUNDANCE 267

25-]

'»H

15 0 1. 8 0. 6 0. 4 0. 2 0. 0 0. r

1000- K (millions)

MSY (millions) 100^

0.0 0.2 0.4 0.6 0.8 1.0

FIGURE 10.—Population growth rate (r) versus error sum of squares (SSE; upper panel) and versus estimates of maximud Kan m sustainable yield (MSY; lower panel r rivefo ) r herring, base n 1845-190do 7 CPUE data.

vary considerably, but the changes from year to r AmericaFo n shad r adequatbion a , ou -f o t efi year appea refleco t r t onlno ty sampling errot bu r mass-based model to the CPUE data couJd not be als oserialla y correlated process. obtained for the period 1845-1907. We attributed Of much poorer quality are the landing records, difficulte th fittinshare n i y th o pgt declin hauln ei - and we made numerous assumptions and inter- seine CPUE observe n 1874-1876i d . Tha- oc t polation generato st annuae eth l landing estimates. curre t abouda same th t e timreportea s ea d decline The effect of these manipulations is unknown, so in catch rate f Americaso n shad migratine th p gu we have relied on the long history of the fishery Roanoke River to spawn. John Emry, the owner to provide some management guidance- ex r Fo . fise th h f slido t Weldonea , reporte ddecreasa n ei ample, the estimate of MSY for river herring, American shad abundance "o t leasa f t one-haln i f based on the 1845-1993 model, was 5.7 million the last ten years" (Yarrow 1874). Similar declines kilograms. Tha similas i taverage th o t r e reported in CPUE were reported at that time for several landings from 1880 to 1970 (5.4 million kilo- southern river systems, including the Savannah, grams), whic s clearlwa h a ysustainabl e harvest Congaree, Wateree, and Cape Fear rivers (Yarrow under historical condition f stoco s k productivity. 1874). Yarrow (1874) suggested that intensive Thus onle th , y remaining questio whethes ni r hab- fishing throughout the south was responsible for itat has been lost or degraded to such a degree that e declineth abundancen i s e furtheH . r suggested historical level f harves slongeo o n e ar rt possible. tha te fiscontinueth h f o slid t e Weldoa eus d n HIGHTOWE. AL T RE

Haul seine CPUE r2500

Predicted pount dne QPUE

Predicted haul seine CPUE 1993 model Poun CPUt dne E

Predicted haul seine CPUE 1907 model 1845 1855 1865 1875 1885 1895 1905 1915 1925 1935 1945 1955 1965 1975 1985 Year FIGURE 11.—Observed and predicted river herring CPUE for the haul-seine and pound-net fisheries.

would resul n overharvesi t f Americao t n shad effort within a data series. The first is a nearness within a few years; unfortunately, landing data that index (/V), which ranges fro m(leas0 t reliableo t ) could be included in our model were not available. 1 (most reliable) and indicates how closely a mod- Also, we have no recent CPUE data for Amer- eled stoc s approachekha biomase dth s level pro- ican shad because the takee yar n primaril gilly yb - ducin (#MSY)Y gMS t iI s- defines a d netting, rathe poune r th tha y dnb nets- useob o dt tain recent river herring CPUE. The model fitted *MSY yv= i - (6) the historical American shad CPUE data reason- abl 30-yeae y th wel r fo lr period 1845-1874t bu , B* is the smaller of K or the estimated stock bio- e reliabilitth f thayo t mode s difficuli l judgeo t t . estimateY MS e sTh though e mosb o t t feasible mass closes #MSYo t t f biomasI - s estimatei s o dt have crossed ^MSY* the index is defined to equal (0.9-1.8 million kilograms) were greater thae nth e secon1Th . d measur a coverag s i e e index (£), estimated landings during the haul-seine fishery which ranges fro m0 (leas t reliable 2 (mos o t ) t but less than peak landings froe pound-nemth t reliable indicated an ) widelw ho s y stock biomass fishery (which were clearl t sustainable)yno . Better has varied between 0 and K. The coverage index estimate f sustainablo s e e possiblharvesb y ma te is defines a d inden ia f f currenxo t abundanc developee b n eca d for this species. r - B Experience with fitting biomass-based models C = (7) shows ha n tha bese th t t obtainefite ar s d when there #MSY is considerable contrast in stock size and fishing B' is the smaller of K or the largest estimated stock effort (Hilbor Walterd nan s 1992) than I . t regard, biomass, and B is the smallest estimated value. one advantage of our historical data set is the pe- r estimateOu d nearnes r rivefo 0 rs1. inde s wa x riod without fishing during the Civil War. We be- herring, compared to 0.7 for American shad. Es- lieve that the temporary closure was important in timated coverage indices were 1.6 for river herring establishing how resilient the stocks were to fish- and 0.6 for American shad. These values suggest ingr riveFo . r herring alse w , o believe thalone th t g that the river herring results are likely to be more serie f landingo s s valuabli s e becaus t includeei s reliable than thos r Americaefo n shad f courseO . , an extended perio f relativeldo y constant landings these indices are only one indication of reliability; recena d an t rapid declin landingsn ei . for example t tak, theno e o intyd o accoun gape th t s Prage t ale r . (1994) propose quantitativo dtw e in the landings data used to generate our biomass measure contrase th f so stoc n i t k sizfishind ean g estimates. HISTORICAL TREND ALOSAN SI ABUNDANCE 269

SO-, 45- 40- 35- fao- I25-

10 15- 10- 5- 0 8 0. 0.2 6 0. 0.4 r 1000-

"5 100-

10-

0.0 0.2 0.4 0.6 0.8 1.0

FIGURE 12.—Population growth rate (r) versus error sum of squares (SSE; upper panel) and versus estimates of maximud Kan m sustainabic yield (MSY: lower panel river )fo r herring, base 1845-190n do d 1977-1997an 3 CPUE data.

Another question in interpreting the reliability f colonizatioo f previouslno y unavailable spawn- r resulte higou th hf s o i svariabilit n CPUEi y , ing habitat. We considered revising our model to which may result from year-class fluctuations. account for year-class variability (Prager 1993), Standard biomass-based models assume that sys- but we were unable to obtain auxiliary data on tematic changes in CPUE result from the popu- flo temperaturr wo e fro 1800e mth s that might have lation's response to fishing. For that reason, our been related to fluctuations in year-class strength. misinterpretine modeb y ma l g year-class variation We hope that the results of this study will pro- as an indicator of high population growth. That vide an impetus for restoration of the Albemarle type of error seems more likely for American shad Sound stocks of American shad and river herring. (? = 0.5-0.9) than for river herring (/> = 0.3). We These populations supported longstanding tradi- had expected a lower estimated r for American tional fisheries that were onc considerablf eo e eco- shad than for river herring because of the shorter nomic importance to the region. The populations average life spa r rivefo n r herrin n Albemarli g e e estimatear havo dt e relatively rapid growtd han Sound (Winslo t alwe . 1983) r bestOu . estimatf eo were shown to rebound quickly in response to a river fo r r herrin similas gwa estimateo rt w Ne r sfo period without fishing. Restoration programs for BrunswicEnglanw Ne d dan k stocks (range, 0.303- American sharived dan r herring have been initi- 0.54 alewivesr 0fo ; 0.437-0.55 bluebacr 0fo k her- ated in Maine, New Hampshire, Rhode Island, ring: Crecco and Gibson 1990). The Crecco and Massachusetts, Pennsylvania, Delaware, Mary- Gibson (1990) estimates were based on observed land, Virginia Soutd an , h Carolina somd an , e pop- increase population si nfirse sizth t n e8-1i 0 years ulations have increased dramatically (Crecco and 270 HIGHTOWR. AL T RE

Gibson 1990; ASMFC 1993). Measures useo t d marle Sound stocks. For example, on the basis of rebuild those populations include stocking (frd yan current prices, the river herring fishery would in- adults), restriction n harveso s t (including clo- crease in value from less than U.S. $100,000 to sures), construction and operation of fish passage about $1,000,000 if landings could be restored to facilitie t damsa s (including fish lifts) habitad an , t the average level reported for 1880-1970. Al- improvement (ASMFC 1993; Coope t ale r . 1994; though water qualit s generallyha y improved over Rulifson 1994). the past 20 years (Eades and Cooper 1994), the Our results suggest that relatively short periods question of whether that long-term average harvest f vero y high harvest initiate e declinedth e th f o s is sustainable give currene nth t habitat qualit- yre Albemarle Sound stock Americaf so n sha rivd dan - mains to be answered. er herring. For river herring, our results are con- sistent with those from studies linkin decline gth - Acknowledgments ing fisher excessivo t y e offshore catche y forb s - We appreciate the review comments provided eign vessels during the 1970s (Street and Davis by D. Ahrenholz, W. Laney, J. Kornegay, C. Man- 1976; Rulifson et al. 1987; Crecco and Gibson ooch . PragerM , . RulifsonR , . WinslowS d an , . T . 1990). For American shad, harvests were well Taylor and N. Zeller were extremely helpful in above our estimated MSY around 1900, and land- findin datad g ol Taylor'. .T s appreciatio f historno y ings have declined more or less continuously since provide e inspiratiodth papere th r .nfo then. Reports from the late 1800s attributed the disappearanc f vaseo t run f Easo s t Coast anadro- References mous fish to intensive fishing and loss of spawning ASMFC (Atlantic States Marine Fisheries Commission). habitat becaus constructiom da f eo n (Worth 1882; 1993. Atlantic States Marine Fisheries Commission McDonald 1887). Smith (1907) suggested that ex- 1993 shad and river herring workshop, Virginia cessive fishinr nea o e ocea th n r o g n during that Beach, Virginia. ASMFC, Washington, D.C. time seriously threatened the fishery by limiting Boyce. S 1917 . W ,. Economi d sociaan c l historf o y Chowan County, North Carolina, 1880-1915. Co- the supply of fish on the spawning grounds. lumbia University Press Yorkw Ne . . What remains to be determined is the degre . eDavisS t. oH . d Chestnut1975an , F. . . SynopsiA , f o s which current landings affect these stocks. Current marine fisherie f Norto s h Carolina, part I. Univer- harvests are well below historical levels and the sity of North Carolina, Sea Grant Program Publi- estimate exploitatiof e MSYso th t ,bu n rates could cation UNC-SG-75-12, Raleigh. still be excessive because population sizes are Eades. T . Loesch . KlaudaCooper. J G R . . . J ,R d E. . ,,an ,J though lowe b have o t .informatioW to en n about editors. 1994. Anadromous Alosa symposium. American Fisheries Society, Tidewater Chapter- Be . current exploitation rate r populatioo s n sizet bu , thesda, Maryland. our estimated r for river herring (0.3) suggests that Crecco. V. A., and M. Gibson. 1990. Stock assessment the exploitation rate (M) should be low (UMSY = of river herring from selected Atlantic coast rivers. r/2: Hilborn and Walters 1992). For American Atlantic States Marine Fisheries Commission, Spe- shad, our estimates of r (0.5-0.9) correspond to a cial Report 19. Washington. D.C. recommended exploitation rate of 0.25-0.45. Tag- . GibsonM Creccod an . , 1988V , . Method f estimatso - ing fishing mortality rate American so n shad stocks. ging studies could be done to estimate current rates Appendix B in Supplement to the fishery manage- of exploitation (Crecco and Gibson 1988). Those ment pla r thnfo e anadromous alosid stocke th f so studies would need to also address ocean harvest eastern United States: American shad, hickory shad, of American shad rivean d r herring, whics ha h alewife, and blueback herring. Atlantic States Ma- been increasing and, for American shad, represents rine Fisheries Commission, Washington, D.C. a substantial fraction of the total harvest (Harri . CooperE . J s d . an Eades 1994, anadromoue T . Th .R , s and Rulifson 1989). Alosa symposiu n retrospecti m : what have w e learned? Pages 159-16 Cooper. E . Eades. J T 0 n . i .R , firse Th t step toward restoring these stocks could R. Loesch. Klauda.J G . J d , ,an editors . Anadromous asseso t e effece b sth f fishino t g (including ocean Alosa symposium, American Fisheries Society, fisheries) on survival, and the effects of water Tidewater Chapter, Bethesda. Maryland. quality and loss of spawning habitat (including the Earll.R . 1887E . . North Carolin fisheriess it d aan . Pages Roanoke Rapids Dam for American shad) on re- 475-497 1/1 G. B. Goode, editor. The fisheries and production f overfishinI . s occurringi g r estiou , - fishery industries of the United States, section 2. U.S. Commission of Fish and Fisheries. Washing- mate historicae MSf th so d Yan l landings data sug- ton. D.C. gest that there is much to be gained by restricting Harris, P. J.. and R. A. Rulifson. 1989. Investigations harvest over the short term to rebuild the Albe- f oceao n landing r Americasfo n sharived dan r her- HISTORICAL TRENDS IN ALOSA ABUNDANCE 271

ring from United States east coast waters. Atlantic American Fisheries Society. Tidewater Chapter, Be- Slates Marine Fisheries Commission, 89-1PASRH, thesda. Maryland. Washington, D.C. (revised December 1991.) Rulifson, R. A., S. A. McKenna, and M. L. Gallagher. Hilborn . WaltersJ . C , d R..an , 1992. Quantitative fish- 1987. Tagging studies of striped bass and river her- eries stock assessment: choice, dynamics- un d an , f rinFundy o n uppe i gy Ba r. Nova Scotia. North certainty. Chapman and Hall, New York. Carolina Departmen f Naturao t l Resourced an s . KrieteH . W Loesch. . , TravelsteadJr.G G. . ,. J J , . J . E , Community Development, Divisio f Marinno e Fish- Foell, and M. A. Hcnnigar. 1981. Biology and man- eries, Completion report AFC-28-1, Morehead City. agemen f mid-Atlantio t c anadromous fishes unde- rex Schaefer . 1954B . M , . Some aspectdynamice th f o s s tended jurisdiction. Part II: Virginia. Virginia Institute of populations important to the management of of Marine Sciences Special Report 236. commercial marine fisheries. Inter-American Trop- ical Tuna Commission Bulletin 1:27-56. McDonald. M. 1884a. Chart of the river basins of the Smith, H. M. 1907. The fishes of North Carolina, vol- Atlantic slope with curves showing the present lim- um . Nort2 e h Carolina Geologica d Economian l c its of migration of shad as determined by obstruc- Survey, Raleigh. e natura e th rivertion th d n lan i s s limit f theio s r Slat f Norteo h Carolina, Halifax County Superior Court. migration before the erection of dams. Plate 214 in 1907. Roanoke Rapids Power Compan . Roavs y - G. B. Goode, editor. The fisheries and fishery in- noke Navigation and Water Power Company. Ro- dustries of the United States, Section I. U.S. Com- anoke Navigation and Water Power Company, mission of Fish and Fisheries, Washington, D.C. 1882-1913, CR 047.928.2:96-97. (Stale Archives. McDonald, M. 1884b. The shad and the hickory shad, Raleigh, North Carolina). or mattowacca. Pages 594-608 in G. B. Goode, ed- Slevenson . 1899H . e shaC , Th d. fisherie f iho se Al- itor. The fisheries and fishery industries of the Unit- lanlic coast of the United Stales. Pages 101-269 in Statesd e , Sectio . U.SnI . Commissio f Fisnd o han G. M. Bowers. Reporl of ihe U.S. Commission of Fisheries, Washington, D.C. fisheriesd fis . U.Shan 24 n . Commissiopa , f Fisno h McDonald . 1887 M e ,river soundd Th . an s f Norto s h and Fisheries, Washingion, D.C. Carolina. Pages . Goode625-63B . G ,n 7i editor . Slreel . DavisJ . d W.M .an , . 1976. Nole ihn so e river e fisherieTh d fisheran s y industrie e Uniteth f o sd herring fishery of SA 6. International Commission States. Section V. U.S. Commission of Fish and for the Northwest Atlanlic Fisheries, ICNAF Re- Fisheries, Washington, D.C. search document 76/VI/61, Serial number 3848, Milner, J. W. 1880. The propagation and distribution of Morehead Cily, North Carolina. shad in 1878. Pages 611-628 in Report of the Com- Taylor, H. F. 1951. Survey of marine fisheries of North missione r 1878rfo , par . U.S6 t . Commissio Fisf no h Carolina. University of North Carolina Press, Chap- and Fisheries, Washington, D.C. el Hill. Mittertreiner, A., and J. Schnutc. 1985. Simplex: a man- Taylor . 1992T . M , . Seiner congersd san : North Carolina fisheries in the old and new south. North Carolina ual and software package for easy nonlinear param- Historical Review 1:1-33. eter estimatio d interpretatioan n n fisheri n - re y . UArmS . y Corp f Engineersso . 1943. Review report search. Canadian Technical Report of Fisheries and on Roanoke River. U.S. Engineer Office, Norfolk, Aquatic Sciences 1384. Virginia. North Carolina Board of Agriculture. 1896. North Car- University of North Carolina. 1980. Hayes collection olina and its resources. NCBA, Raleigh. 1694-192 8e Souther i/Th i n Historical Collection, North Carolina Boar f Agriculturedo . 1881. September microfilm cdilion. Universily of North Carolina Li- monthly bulletin. NCDA, Raleigh. brary, Chapel Hill. . Prager1993H . suitA M . , f extensiono e a non o t s - Winslow . 1990E . S , . Statu f Americao s n shad, Alosa equilibrium surplus-production model. U.S. Nation- sapidissima (Wilson) Nortn i , h Carolina. North Car- al Marine Fisheries Service Fishery Bulletin 92: olina Departmen f Environmento t . Health Natd an , - 374-389. ural Resources, Division of Marine Fisheries, Spe- Prager, M. H. 1994. User's manual foi ASPIC: a stock- cial Scientific Report 52, Morehead City. production model incorporating covariates, pro- Winslow, S. E., and six coauthors. 1983. North Carolina gram version 3.5x. U.S. National Marine Fisheries anadromous fisheries management program. North Service, Miami Laboratory, MIA-92-/93-55, Mi- Carolina Department of Environment. Health, and ami, Florida. Natural Resources, Divisio f Marino n e Fisheries, Prager. M. H.. C. P. Goodyear, and G. P. Scott. 1994. Complelion Report AFCS-16. Morehead Cily. Application of a stock-production model to age- Worth, S.G.I 882. Report of superinlendenl of fish and structured simulation data: a swordfish-Iike stock. fisheries r ihefo , years 1879-'80. Pages 25-5n 1i ReporStale th f e o tBoar f Agriculturdo r 1881efo , International Commission for the Conservation of session 1881, par . Slal2 t e Boar f Agriculiureo d , Atlantic Tunas, Working document SCRS/94/116 Raleigh, North Carolina. (revised), Madrid. Yarrow . 1874C . H , . Repor reconnaissanca f o t f theo e Rulifson, R.A.I 994. Statu anadromouf so s Alosa along shad—rivers south of the Poiomac. Pages 396-402 the east coast of North America. Pages 134-159 in i/i Repor f iho t e Commissione r 187 fo rd 1873 2an , . CooperE . Eades. . J T . G Klauda. J . . R ,J R . d an , part 2. U.S. Commission of Fish and Fisheries, Loesch, editors. Anadromous Alosa symposium. Washingion, D.C.