Nitrogen Content and Other Soil Properties Related to Age of Red Alder Stands1

B. T. BORMANN AND D. S. DEBELL2

ABSTRACTThe magnitude and pattern of nitrogen (N) accretion and

other changes in soil properties were assessed for red alderstands (Alnus rubra Bong.) 5 to 41 years old growing on thesame soil type in the same general area. Regression of soilnitrogen (X) content over stand age revealed that N has ac-cumulated at a nearly constant rate of about 35 kg ha"1 year'1in the mineral soil (0- to 20 cm depth) beneath alder stands.After an apparently rapid build-up in the first decade, forestfloor N increased linearly from 10 to 40 yearsi at a rate of15 kg ha'1 year"1. Other mineral soil characteristics beneathalder stands differed markedly from those beneath adjacentDouglas-fir stands [Pseudotsuga menziesii (Mirb.) Franco]; or-ganic matter content was 20% higher, and pH and bulk densitywere much lower. The rate of N accretion and improvementof other soil characteristics suggest opportunities for increasingyields of Douglas-fir grown in mixed or rotation;]! culture withred alder.

Additional Index Words: Alnus rubra, N accretion, soil de-velopment, Douglas-fir.

Bormann, B. T., and D. S. DeBell. 1981. Nitrogen content andother soil properties related to age of red alder stands. SoilSci. Soc. Am. J. 45:428-432.

RED ALDER (Alnus rubra Bong.) is the major hard-wood in coastal forests of the Pacific Northwest.

Its rapid occupation of disturbed forest land and fastearly growth make it appear as an unwelcome com-petitor of the highly valued Douglas-fir (Pseudotsugamenziesii Mirb. 'Franco'). However, its effects on thesite may enhance Douglas-fir yield. Tarrant (1961)found that the current diameter and height growth of30-year-old Douglas-fir trees interplantecl with alderwas better than the diameter and height growth inan adjacent pure plantation of Douglas-fir of the sameage. Fifteen years later, Miller and Murray (1978)evaluated the same two stands and found even greaterdifferences; total average volume of the mixed standwas 159 m3/ha (Douglas-fir 88 m3/ha and alder 71m3/ha). In contrast, total average volume of the purestand of Douglas-fir was only 82 m3/ha. The mixedstand contained 938 kg/ha more soil nitrogen (N)and 31% more soil organic matter than the pure stand(Tarrant and Miller, 1963). Substantial increases insoil N have also been observed in other red alderstands (Tarrant et al., 1969; Newton et al., 1968; Bergand Doerksen, 1975).

Because N is a major limiting factor to growth ofDouglas-fir (Gessel et al., 1969; Miller and Fight,1979), Tarrant and Trappe (1971) have proposed thatthe ability of red alder to fix atmospheric nitrogen(N2) be exploited by growing the species in alternaterotation with Douglas-fir. Growth estimates for man-aged alder plantations indicate that pulp logs can be

1 Paper no. 1460 of the Forest Research Lab., School of For-estry, Oregon State Univ., Corvallis, OR 97331. Received 8 Apr.1980. Approved 8 Oct. 1980.

2 Research Assistant, Dep. of Forest Science, School of For-estry, Oregon State Univ., Corvallis, and Principal Silviculturist,USDA Forestry Sciences Lab., Pacific Northwest Forest & RangeExp. Stn., Olympia, Wash., respectively.

produced in 10 to 15 years and sawlogs in < 40 years(DeBell et al., 1978). But to begin to evaluate croprotation options, we need information on the relationof N accumulation to the age of red alder stands.

The present study was designed to help meet thisneed and was conducted on sites that are highly pro-ductive for alder growth, thus on sites where opportu-nities for crop rotation are greatest. Also investigatedwere changes in soil pH, organic matter, and bulkdensity.

PROCEDURESDescription of the Study Area

The forest stands studied are within an area 4 km in radiusin the Capitol Forest near Olympia, Wash. Elevation is 200to 300 m above sea level, and climate is characterized by warm,dry summers and cool, wet weather during the rest of the year.Mean annual precipitation exceeds 1,500 mm. All plots werelocated on deep soils of the Boistfort series developed fromEocene basalt on level to moderately steep upland terraces.Douglas-fir productivity here is high, with site index averaging165 feet (50 m) at 100 years (McMurphy and Anderson, 1968).

Before 1875, the Capitol forest area was covered with old-growth conifer forests of Douglas-fir, western hemlock [Tsugaheterophylla (Raf.) Sarg.], and western redcedar (Thuja plicataDonn) and presumably only small amounts of red alder alongstreams. Most old growth had been logged by the early 1940'sand much of the area was planted with Douglas-fir. SomeDouglas-fir and other conifer and hardwood species have regen-erated naturally. Present overstory vegetation consists pri-marily of a mosaic of pure and mixed stands of Douglas-firand red alder, with a scattering of other conifer and hardwoodspecies. Common understory species now include swordfern[Polystichum munitum (Kaulf.) Presl. var. munitum], Oregon-grape (Berberis aquifolhim Pursh), elderberry (Sambucus sp.)tsalal (Gaultheria shallon Pursh), vine maple (Acer circinatumPursh), and salmonberry (Rubus spectabilis Pursh).

General ApproachThe age-sequence method was used to assess patterns of N

accretion and other changes in soils beneath alder, measuringsoil characteristics in stands of different ages but of presumablysimilar origin (e.g., Newton et al., 1968; van Cleve et al., 1971;Silvester, 1977). Because of concern about natural variabilityin soil N content, we also used the paired-plot method, com-paring adjacent stands that presumably differ only in speciescomposition (e.g., Tarrant and Miller, 1963; Berg and Doerk-sen, 1975; Cole et al. 1978).

Plot SelectionCandidate plots were those in pure red alder or Douglas-fir

stands meeting the following criteria: (i) Boistfort soil series;(ii) sufficient size to eliminate edge effects on the samplingarea; (iii) full, uniform stocking; (iv) slope < 20%; and (v)evidence (large stumps) that original vegetation was old-growthconifers. Thirteen red alder plots were selected for good agedistribution and similarity in other characteristics such as as-pect. Proximity to a Douglas-fir stand for paired-plot compari-sons was also a selection factor. Six Douglas-fir stands wereselected; some of these were paired with more than one alderstand to provide 10 paired-plot comparisons.

Field and Laboratory MethodsTwo perpendicular 10-m transects, crossing in the center,

were laid in north-south and east-west directions. Forest floorand mineral soil samples were collected along these transects.Average stand age was determined from increment cores ex-tracted breast high from all trees located in a 50-m2 (7.07 by

428

BORMANN & DE BELL: N CONTENT AND SOIL PROPERTIES RELATED TO AGE OF RED ALDER STANDS 429

Table 1—Soil characteristics of alder and Douglas-fir stands.

Mineral soil (0-20 cm)

Plot

ABCDEFGHIJKLMMean

1-ET2-H.M3-F.L4-J5-A.G.K6-1

StandageYear

5.29.59.8

11.513.724.427.028.729.233.537.038.040.623.7

26.030.533.333.538.041.5

Forest floor

Dry weight

Metric ton/ha

14.2

20.613.513.138.5

35.3

39.325.0

14.324.424.8

21.6

Nitrogen

%

2.0

1.92.11.91.5

1.8

1.81.9

1.11.11.0

1.0

kg/ha

290

380280250590

650

730450

150260260

220

Bulkdensity

g/cm3

Alder0.740.600.530.490.750.770.690.770.400.890.770.920.690.69

Douglas-fir0.780.850.900.930.850.69

2-mm fraction

Nitrogen

%

0.260.270.410.230.210.290.330.290.440.250.310.250.360.30

0.220.210.190.190.220.22

kg/ha

2,5703,0603,1702,1102,8003,5103,5203,3503,3103,8803,3603,5903,9503,240

2,7802,7402,7502,8802,8002,870

Organic matter

%

9.28.6

18.47.47.58.1

10.09.0

14.06.58.86.5

11.89.7

6.86.55.75.97.26.8

Metric ton/ha

10096

14467

10097

1051111061029693

128104

808582919188

PH

5.25.24.75.85.45.45.04.94.85.35.05.34.05.1

5.85.96.86.25.75.6

t Letters refer to paired alder stand.

7.07 m) quadrat formed by stakes at each end of the transectlines. *

The forest floor (litter layer) was sampled with a template(960 cm2) at 40 points spaced 0.5 m apart along transect linesand composited for each of seven red alder and four Douglas-firstands. Fresh weight of the composite was determined in thefield. A subsample was removed and weighed, oven-dried at75°C, and reweighed to estimate dry weight of forest floor perhectare. The subsamples were then ground in a Waring blenderand analyzed for total N by macro-Kjeldahl procedures (Jack-son, 1958).

Mineral soil was collected by a tube sampler (4.8 cm i.d.),which yielded 362 cm3 of soil per sample. All data reportedhere came from 0- to 20-cm depth. The soil was sampled at 20points spaced 1 m apart along the sampling transects and com-posited for each of the 13 red alder and 6 Douglas-fir stands.Weight of the fresh composite was determined, and subsam-ples were taken for determining pH and moisture content. Soilacidity was measured with a pH meter in 1:1 soil/water paste.Data for fresh and oven-dry (105°F) weight determined on thesubsamples were combined with data for volume and compositewet weight to calculate bulk density. The remaining portionsof the composite samples were air-dried before being storedfor about 6 months. A subsample of the air-dried compositewas rolled and sieved to yield a < 2-mm fraction which wasground in a Bico pulverizer. From this, total N was determinedby a macro-Kjeldahl technique modified to include nitrates(Chapman and Pratt, 1961). Soil organic matter was determinedby the Walkley-Black procedure (Jackson, 1958).

Statistical Analysis

Relationships between alder stand age and individual soilvariables were analyzed with linear regression. Differences be-tween certain soil characteristics in alder and Douglas-fir standswere examined with paired and unpaired t-tests. Multiple re-gression techniques were used to determine relationships amongseveral variables.

RESULTS AND DISCUSSION

Forest-floor and mineral-soil characteristics of eachred alder and Douglas-fir plot are summarized in Ta-ble 1. Here we will examine the relationships betweenstand age and various soil characteristics individually.

Forest Floor Dry Weight and NitrogenThe forest floor weight was assumed to be negligible

at age 0 because dense stands of alder rarely becomeestablished except on exposed mineral soil. Moreover,no conifer litter was observed in any of the sampledalder stands. Given this assumption, total dry weightof the forest floor beneath alder apparently increasesrapidly for the first 5 years. It then accumulates moreslowly to about age 25, when it may reach equilibrium(Fig. 1). Though we cannot firmly establish whenequilibrium is reached with this data, it appears tobe later than the 6 years predicted by Zavitkovski andNewton (1971). Their data include only alder litter,whereas ours include the contribution of all plantspecies, and field observations indicate that litterfallfrom understory species is appreciable in older stands.

§ 40

i30O5 20

10

400

200

o o y = 8.3-K>.87(x)r = 0.800p<0.02

= 174+14.6<x)r = 0.928p<0.01

10 20 30STAND AGE (yr)

40

Fig. 1—Accretion of dry weight and N weight of the forestfloor beneath red alder.

430 SOIL SCI. SOC. AM. J., VOL. 45, 1981

The accretion of N in the forest floor follows a pat-tern similar to that of total dry weight, but the ten-dency toward equilibrium appears less strong (Fig. 1).By age 10, the litter layer contained about 300 kgN/ha; accretion rate from age 10 to age 40, predictedby the slope of the least squares line, is about 15 kgN/ha annually.

Mineral SoilBULK DENSITY

Average density of mineral soil (0-20 cm) under redalder stands was 0.69 g/cm3, 16% lower (p < 0.01,Mest) than in adjacent Douglas-fir stands (0.83 g/cm3).This parallels differences found by Tarrant and Miller(1963) between soil density beneath a mixed red alderand Douglas-fir plantation and an adjacent pure Doug-las-fir plantation. Contrary to expectation, no signifi-cant relationship appeared between bulk density andstand age (Fig. 2).

NITROGENNitrogen concentration in soil beneath alder stands

averaged 0.30%, 41% higher (p < 0.01, Mest) than insoil beneath Douglas-fir (0.21%). When N contentwas expressed in kg/ha, differences were significantbut not as pronounced, because soil bulk densitiesunder alder were lower than under Douglas-fir. Ni-trogen in mineral soil averaged 3,240 kg/ha for alderstands, 16% more than the 2,800 kg/ha for Douglas-fir stands. Similar findings of increased soil N in thepresence of red alder have been reported by Tar-rant and Miller (1963), Tarrant et al. (1969), andBerg and Doerksen (1975). Total amounts of soil Nwere significantly related (R - 0.802, p < 0.01) to

1.0

s-o.9

I0-7

uia 0.6

0.5

0.4

«-DOUGLAS-FIR MEAN

o o

oo

O O

o

- DOUGLAS-FIR MEANO

o

o

o oo

10 20 30STAND AGE (yr)

40

stand age (Fig. 3A). Apparently N is accumulatingin the 0- to 20-cm soil horizon of alder stands at arate of about 34 kg ha"1 year"1; moreover, the rateappears to remain constant throughout the range ofstand ages sampled (5-40 years). This finding conflictswith the hypothesis of Newton et al. (1968) that N2fixation in alder is rapid at first but insignificant af-ter 20 years.

When differences in N content between adjacentalder and Douglas-fir plots were regressed against alderstand age, we obtained slope and correlation coeffi-cients (Fig. 3B) similar to those we have described.Thus, estimates derived by the two approaches confirmeach other. The two techniques show an average an-nual N accretion in the mineral soil of about 35 kg/ha.

The negative y-intercept of the paired-plot regres-sion, was unanticipated and somewhat disconcerting.We suspect that many alder stands have become estab-lished on areas lower in soil N content than areas nowsupporting Douglas-fir stands. Greater amounts ofcharcoal were found in the mineral soil of alder standsthan in soils beneath Douglas-fir stands (p = 0.01,paired Mest), which may indicate N depletion by fire.The natural tendency for alder to become establishedon bare mineral soil and on subsoil (e.g., erodedareas) also supports the hypothesis that soil N contentsat time of stand establishment were lower for alderthan for Douglas-fir.

Estimated rates of annual N accumulation in themineral soil for this study are somewhat lower thanthose reported by others. Berg and Doerksen (1975)estimated an annual accretion rate of 45 kg/ha forunderstory alder, and Tarrant and Miller (1963) esti-mated 40 kg/ha for a mixed alder-Douglas-fir planta-tion. Soils in these studies initially contained muchless N than the Boistfort soils of our study.

The combined rate of N accretion for the forestfloor and mineral soil is about 50 kg ha"1 year"1 from10 to 40 years. A more rapid build-up probably occurs

4000

3000

a0)2000£

J1200z

800

400

-400 It

DOUOLAS-FIR4-MEAN

rs 0.802p<0.01

+ 33.5(x)

10 20STAND AGE (yr)

30 40

Fig. 2—Bulk density and pH of mineral soil (0-20 cm) beneathred alder.

Fig. 3—Nitrogen accretion in the < 2-min fraction of mineralsoil (0-20 cm) beneath red alder. Red alder N weight vs.alder age (A). Red alder N weight minus Douglas-fir N weightvs. alder age (paired plot technique) (B).

BORMANN & DE BELL: N CONTENT AND SOIL PROPERTIES RELATED TO AGE OF RED ALDER STANDS 431

2500

20QO

J1500

Z01000

500<2mm MINERAL SOIL (O-2Ocm)

10 20STAND AGE (yr)

30 40

Fig. 4—Nitrogen accretion in the forest floor and < 2-mmfraction of the mineral soil (0-20 cm) beneath red alder.

in the first decade (Fig. 4). Other data collected inour plots suggest that considerable amounts of N areaccumulating in the lower horizons (20-50 cm) ofthese deep, well-drained soils—perhaps as much as 30kg ha"1 year"1.3 This, in addition to accumulation ofN in biomass, suggests an annual N accretion rate inthe ecosystem of over 100 kg/ha.

Annual accretion rates suggested by Newton et al.(1968) for stands of 14 years or less were more than300 kg N/ha. Our lower values may reflect the effectof higher N content of soils sampled in our studycompared with those of Newton et al. (1968). Otherinvestigations with legumes (Gibson, 1977) and alder(Zavitkovski and Newton, 1968) have shown that N2fixation can be reduced by high levels of N in the soil.

SOIL ORGANIC MATTERSoil organic matter content was about 20% higher

(p < 0.01, (-test) under alder stands than under Doug-las-fir. Increased soil organic matter under alder wasreported earlier by Tarrant and Miller (1963). Be-cause the content is quite variable, especially in theyounger stands sampled, no significant trends withage were evident (Fig. 5). Reasons for a high vari-ability in organic matter than in soil N are unknown.Perhaps factors other than the presence of alder, suchas microsite effects on decomposition and incorpora-tion, degree of burning, or age of the previous coni-fer stand, are significantly related to present amountsof organic matter.

The accumulation of soil organic matter beneathalder aids long-term productivity. Organic matter ispositively related to cation exchange capacity and mois-ture retention. Moreover, it can provide a stable stor-age mechanism for retaining the N fixed by alder untilit can be used by subsequent crops. In addition, Har-vey et al. (1979) showed that increased soil organicmatter leads to greater ectomycorrhizae development.

SOIL ACIDITYThe pH of mineral soil (0-20 cm) beneath red alder

was significantly lower (p < 0.01, Mest) than pH ofsoil under adjacent Douglas-fir stands. This findingagrees with previous reports of increased acidity inred alder stands (Tarrant, 1961; Bollen et al., 1967;Franklin et al., 1968). Although there is a tendency

3B. T. Bormann. 1978. The magnitude and pattern of nitro-gen accretion in Alnus rubra stands. M.S. Thesis. Univ. ofWashington.

160"if§140

£ 120III

8"5 soo« 60

- DOUGLAS-FIR MEAN

O

10 20 30STAND AGE (yr)

40

Fig. 5—Organic matter content of the < 2-mm fraction of themineral soil (0-20 cm) beneath red alder.

for pH to decrease with increasing stand age (Fig. 2),this trend is not significant (p — 0.15). Increased de-composition rates, greater organic acid production, andhigher nitrification rates under red alder may accountfor the difference in soil pH between alder and Doug-las-fir stands and also for the tendency of pH to de-crease with increasing stand age. Whether the lowerpH of soil beneath alder stands will be detrimental toproductivity over long periods is unknown, but it maycontribute to the lower base status and decreased avail-ability of other nutrients reported for alder soils (Bol-len et al., 1967; Franklin et al., 1968).

Simple linear regressions among alder soil variables(Table 2) suggest that soil pH is strongly related toorganic matter and moderately related to soil Nweight. Soil N and soil organic matter, however, wereonly weakly correlated (R — 0.50); such variability inC/N ratio may be a further indication that somefactors other than age of alder stands have importantlyaffected soil organic matter content.

Multiple regression analysis of soil variables indi-cates that together alder stand age and soil organicmatter account for 81.4% of the variability in soil Ncontent and constitute the best combination of vari-ables for estimating soil N content by the Cp criterion(Neter and Wasserman, 1974). The least squares re-sponse surface is:

Y = 1,258 + 31.6(XO + 11.9(X2)where Xi is stand age in years and X2 is soil organicmatter content in metric tons per hectare. Partial co-efficients of determination suggest that stand age ac-

Table 2—Simple linear regressions among alder soil variables.

Regression (Y vs. X) REquation

P-value (Y=b, + b,i

Soil organic matter (metricton/ha) vs. stand age (year) 0.11 >0.20

Nitrogen weight (kg/ha) vs.stand age (year) 0.80 < 0.01 y = 2,450 + 33.5 X

Nitrogen weight (kg/ha) vs.soil organic matter 0.50 0.08 Y = 3,132 + 11AX

pH vs. stand age (years) -0.43 0.15Soil organic matter (metric

ton/ha) vs. pH -0.82 <0.01 Y =272- 33.3 XNitrogen weight (kg/ha) vs. pH -0.61 0.03 y= 6,829 - 707.8 XBulk density (g/cm3) vs.

stand age (years) 0.44 0.14Bulk density (g/cm8) vs. soil

organic matter (metricton/ha) -0.07 >0.20

Bulk density (g/cm3) vs.N weight (kg/ha) 0.46 0.12

Bulk density (g/cm3) vs. pH 0.13 >0.20

432 SOIL SCI. SOC. AM. J., VOL. 45, 1981

counts for 64% and soil organic matter 17% of thevariation in soil N content.

IMPLICATIONSOur analyses indicate that N is accumulating at fair-

ly rapid rates in the forest floor and upper mineralsoil of the red alder stands. At the end of a 10-yearpulpwood or 32-year sawlog rotation (DeBell et al.,1978), we would expect N addition to soil on thissite to be 670 and 1,710 kg/ha, respectively. Most ofthis, and at least some of the N accumulated in biomassor transported to deeper soil layers, would eventuallybe available to a successive crop. Increased soil or-ganic matter should improve soil tilth and stabilizesoil N. Decreased bulk density should also be benefi-cial, but the effect of the trend toward lower pH isunknown.

Understanding of the effects of alder-derived N andother changes in soil properties on site productivitywill help determine the extent of future alder use.Preliminary assessments of economic feasibility ofmixed and rotational cultures of alder in Dowglas-firmanaged forests have been made (Atkinson and Hamil-ton, 1978; Atkinson et al., 1979), but further work isneeded in light of rapidly increasing energy costs andprojected shortages of oil and natural gas. We believethat plantations of red alder or of admixtures of alderand other tree species will become, on some sites, aneconomically feasible alternative to N fertilizer de-rived from fossil fuels. Our optimism is based on thefact that alder can fix appreciable amounts of N on awide range of sites varying from the poorly productiveand low N soils studied by Tarrant and Miller (1963)and DeBell and Radwan (1979) to the productive soilswith high N content sampled in this study.

ACKNOWLEDGMENTWe would like to thank R. W. Walker for his guidance and

insight, T. P. Bormann for her assistance in field sampling, andD. W. Cole for use of University of Washington, College of For-estry Resources analytical equipment. This research was spon-sored by the USDA Forest Service, Pacific Northwest Forest andRange Experiment Station.