in each of five stands per community. The plots were clipped at ground level Aspen Invasion in late July, sorted to grass and grass- likes, forbs, and current annual growth of shrubs, air-dried, and weighed. The in a Portion cover of brush along each survey line was recorded by land surveyors during the original Dominion of Canada legal of the Alberta Parklands land survey in 1907. Forty samples, each 1 mile in length, were randomly selected from the legal land survey field notes of Cautley (1907) for ARTHUR W. BAILEY AND ROBERT A. WROE study. The same lines were then located on aerial photographs taken in Highlight: Brush areas on ranges in the parkland region of southcentral 1966. For each sample transect, the Alberta have increased from 4.8 to 8.0% of the land area between 1907 and length of the l-mile line covered by 1966. The invasion was not consfanf each year, but was concentrated in two brush in 1907 and 1966 was measured. major periods between 1937 and 1970. Annual herbage production under aspen The two figures were compared to and willow was reduced by 80 to 90% when compared wifh the production of determine whether brush cover had adjacent rough fescue grassland. The invasion of aspen info grasslands was corre- changed during the 59.year period. lated with high remperorures, particularly during the monrh of June, I and 2 Four aspen groves were randomly years before tree establishment and with low precipitation 2 years prior fo selected to estimate tree age. The small esrablishment. trees (young stands) on the outer perimeter of each grove were sampled The aspen parkland is a major moderate to strongly rolling, known separately from the taller, larger vegetation zone where groves of aspen locally as knob and kettle topography. diameter trees (mature stands) near the center of the grove. In 1970, 40 (Populus rremuloides Michx.) and wil- The dark brown soils had developed aspen trees in each grove, 20 in the low (Salix spp.) are interspersed with on glacial till. Average annual precipi- young stand and 20 in the mature grassland. The parkland zone is a tation was 36 to 41 cm (14 to 16 stand were randomly selected, cut tension area between the prairie to the inches). The frost-free period was 100 dawn, the height measured and a to 110 days. The July and January south and the boreal forest to the cross-section of each stem was col- north. The prairie vegetation occurring mean temperatures were 18” C (64°F) lected. During the winter, the block within the parklands has been de- and -13°C (9’F), respectively. The from each cross-section was sanded scribed regionally as the rough fescue study area had been lightly grazed and the growth rings were counted (Festuca scabreNa Torr.) and the since white settlement. using a 7 to 20X binocular micro- mixed prairie associations of western scope. Since most of the trees were Methods Canada (Coupland, 1950). less than 35 years old, false rings were During 1969, available annual pro- either not present or were easily recog- The invasion of fescue grassland by duction in aspen, willow, and rough nized. brush (Fig. 1) has been recorded by fescue communities was measured by Correlation coefficients and regres- Moss and Campbell (1947) and randomly locating two 0.25~m2 plots sion equations were developed relating Johnston and Smoliak (1968) in Alberta, by Coupland and Maini (1959) and Maini (1960) in Saskatche- wan, and by Blood (1966) in Mani- toba. The purpose of our study was to establish whether brush invasion had occurred in southcentral Alberta and to determine the rate of invasion and the effect of brush on herbage produc- tion.

The study area was located on the southern edge of the parklands of southcentral Alberta in Township 34, Range 19, west of the 4th Meridian, about twenty-five miles south of Stettler, Alta. The topography was

The authors are associate PrOfeSSOr of range ecology, University of Alberta, Ed- nlont0n, and range survey coordinator, Al- berta Department of Lands and Forests, Lethbridge. The junior author is at present range management specialist, Alberta De- partment of Agriculture, Hanna, Alberta. Manuscript received May 10,1973. tree establishment to environmental factors. The precipitation data used were the average of data from the .P.Pi,l three closest meteorological stations of the Canada Department of Transport, Meteorological Division. Temperature k! ih data were obtained from the meteoro- . --. logical station at Calgary, Alta. Prairie

Results Vegetation production under the closed canopies of the willow and aspen communities was reduced sig- nificantly compared to the adjacent grassland (Table 1). The grass and grass-like production was reduced to 310 kg/ha under willow and 154 kg/ha under aspen compared to 1944 kg/ha in the adjacent grassland. This repre- sents a reduction of grass and grass- likes of 84% under willow and 92% brur .. ._ under aspen. ..I Most one-mile transects had some . biu, brush on them in 1907 and 1966, but the mean area occupied by brush increased from 4.8% to 8.0% (Table 2). Of the 40 one-mile transects sampled, 68% increased in brush cover while only 7% decreased. The net wllw._ increase in brush cover was 3.2% ova 59 years; the annual rate of invasion averaged 0.08%. brush In 1907 the brush cover for the one-mile transect presented in Fig. 2 was restricted to five stands of willow. By 1966, the same line crossed five stands of aspen and eight stands of willow. Cautley (1907) recorded only two groves of aspen crossing the 40 one-mile transects. In 1966, 35 of the 40 transects (88%) had one or more aspen groves present. The growth ring analysis indicated major periods of rapid tree invasion o YOUNG STANDS (Fig. 3). Two periods, 1960 to 1965 (5 w MATURE STANDS to 10 years ago) and 1944 to 1950 (20 to 26 years ago) had the highest frequency of tree establishment. Of the 160 aspen trees studied, the youngest tree was 3 years old and the oldest was approximately 67 years old (67 growth rings). As aspen grows older, more false rings seem to occur, making estimation of tree age more difficult. Trees in the young aspen stands had diameters ranging from 1.3 to 7.1 cm (0.5 to 2.8 inches); the trees ranged from 1.4 to 8.7 m (5 to 28 ft) tall. In I III ILmI III the mature aspen stands, the tree IO 15 20 25 30 35 40 45 50 55 60 65 diameters ranged from 4.6 to 30.5 cm NUMBER OF GROWTH RINGS (1.8 to 12 inches); the trees ranged from 4.5 to 15.7 m (15 to 51 ft) tall.

264 Table 1. Annual herbage production (kg/ha)’ of willow, aspen, and adjacent rough fescue Alberta by Moss and Campbell (1947) grassland communities in 1969. and Johnston and Smoliak (1968). The rate of invasion of brush on to Willow Aspen Rough fescue grassland, however, averaged only Vegetation class (n = 10) (n = 10) (n = 10) 0.05% per year compared to 0.75% per 1944 Grass and grass-likes 310 154 year in the Porcupine Hills of South- Forbs 218 82 46 Shrubs 0 282 22 western Alberta, an area of higher annual precipitation. The dark brown Total available production’ 528 + 643 518 + 76 20125 88 soil zone in Saskatchewan where Maini ‘To convert kg/ha to lb/acre multiply by 0.89. (1960) worked had a climate similar ‘Total annual production of the rough fescue grassland but only available annual production in the other two communities because annual growth of tall willow and aspen beyond the to that of the study area. He reported reach of livestock was not measured. rates of invasion of 1 .O to 2.5 ft/mile 3 Standard error of the mean. per year; these rates (0.02 to 0.5% per year) are very similar to the average Maini (1960) reported that sucker- which accounted for 41% of the total rate of invasion of 0.05% annually ing was stimulated when temperature variation: found in the study area. and moisture conditions were favor- y= -31.501 - 0.453X, +0.587X2 +0.373X, The area now dominated by aspen able. In this study, the correlation for may have produced more than 1990 all 160 trees between rate of tree kg/ha prior to the invasion of aspen. establishment and annual precipitation where Y = number of trees established, Under mature aspen stands sampled, was nonsignificant (Y 0.02). This value X1 = annual precipitation 2 years 60% of the soils were black cherno- may have been influenced by climatic before tree establishment, X, = mean zems, and the remaining were carbo- extremes. Aspen is a short-lived tree annual temperature 1 year before tree establishment, and X, = mean temper- nated regosolic humic gleysols. In con- detrimentally affected by drought ature for June the year before tree trast, 80% of the soils under rough (Maini, 1960). Few trees were more establishment. The multiple correla- fescue grasslands were dark brown than 32 years old (Fig. 3). The sample tion coefficient (R = 0.642) was signif- chernozems, soils which have of trees surviving prolonged droughts icant (P < 0.01). The t-test on the developed under drier climatic condi- of the 1930’s may be too small to give slope b for each variable (Xl, X, , X,) tions than those under aspen. Aspen a significant correlation. To test the was also significant (P < 0.05). groves generally occur downslope from relationship of tree establishment to A single variable, June mean tem- fescue grasslands or on north-facing temperature and moisture conditions perature 1 year before tree establish- slopes. In the dark brown soil zone of since the drought of the 1930’s, cor- ment was found to be highly signifi- Saskatchewan, Ayyad and Dix (1964) relation coefficients were developed cant (Y = 0.556) for the 22 years have demonstrated that north-facing for trees established in the 1937 to having three or more trees established slopes usually have higher soil mois- 1970 period. Mean air temperatures (Fig. 3). The linear regression equation ture than middle or upper slope posi- and total precipitation were tested for was : tions. An 88% reduction in annual the year of tree establishment as well herbage production is probably a con- Y = -27.861 + 0.607 X as 1, 2 and 3 years prior to tree servative estimate because aspen now establishment for May, June, July, where Y = number of trees established occupies more productive soils. August, September, June to August, and X = mean June temperature the The invasion of aspen into adjacent May to September, and annually. year before tree establishment. This grasslands as has occurred in the young Mean air temperature 1 year prior to accounted for 31% of the variation stands studied was assumed to have tree establishment was significantly and the t-test on the slope b was occurred in years of high precipitation correlated for June (Y = 0.42*, significant (P < 0.01). as indicated by Maini (1960). It has P

Item Multiple regression equations were 1907 1966 developed using 33 years of tree es- Transects having some brush cover (%). 92 95 tablishment data and 51 temperature Transects having no brush cover (%). 18 5 Range in brush cover per mile (%). and precipitation variables. One equa- O-18 O-26 Mean brush cover (%).I 4.8 8.0**’ tion having nine independent vari- Transects that increased in brush cover (%). 68 ables accounted for 55% of the total Transects that decreased in brush cover (%). 7 variation. A more readily understand- Transects that remained the same (%). 25 able equation is the one following ’ Significantly greater (p < 0.01) than brush cover in 1970 using the paired I-test.

JOURNAL OF RANGE MANAGEMENT 27(4), July 1974 265 ment in aspen was found to be nega- Summary and Conclusions environmental complex on prairie slopes tively related to the precipitation and in Saskatchewan. Ecol. Monogr. Annual herbage production under positively related to the mean temper- 34:421-442. brush was only 10 to 25% of that of ature regimes that occurred 1 and 2 Blood, D. A. 1966. The Festuca scabrella the adjacent grassland. Brush cover of association in Riding Mountain National years previous. It would appear that aspen and willow showed a net in- Park, Manitoba. Can. Field Naturalist aspen suckering and invasion of grass- crease of 3.2% during the 59-year 80:24-32. lands is closely linked to high soil Cuatley, R. H. 1907. Field notes of town- period of 1907 to 1966. Age distribu- temperature (Shirley, 1932; Maini, ship 34, range 19, west of the 4th tion of the trees indicated that the Meridian. Dominion of Canada, legal 1968) low precipitation, and possibly invasion was concentrated into two land survey, Field Book 9874, Prov. also to apical dominance (Farmer, major periods of expansion. The in- Dep. Highways Bldg., Edmonton, Alta. 1962). Our findings are in agreement Coupland, R. T. 1950. Ecology of the crease in cover of brush included the with Maini (1968) who maintains that mixed prairie in Canada. Ecol. Monogr. establishment of new stands as well as any disturbance resulting in an 20:271-315. the increase in size of established Coupland, R. T., and J. S. Maini. 1959. increase in soil temperature stimulates stands. In 1907, only two groves of Woodland invasion of western Canadian aspen suckering. The greatest invasion grasslands. IX Inter. Bot. Cong. Proc. II, aspen were recorded in 40 one-mile of aspen into grassland in the study IIA. Montreal. transects but by 1966, 88% of the area occurred in 1962, 1 year after the Diller, 0. 0. 1935. The relation of tempera- transects had aspen groves present. 196 1 drought, when the June mean ture and precipitation to the growth of The rate of invasion may well be beech in northern Indiana. Ecology temperature was 8” F above the 33- increasing because of the greater num- 16:72-81. year average. Suckering continued for ber of brush patches becoming es- Dix, R. L., id 0. A. Steen. 1968. Aspen 1963 and 1964, when the June mean dieback in the parklands of Saskatche- tablished. Over many years, the inva- temperature the year before had also wan. Late Spring Issue, Bull. Ecol. Sot. sion of aspen and willow would mean been higher than average. Amer. p. 81. a loss in forage production and a Farmer, R. E., Jr. 1962. Aspen root sucker Dix and Steen (1968) recorded the decline in carrying capacity of these formation and apical dominance. Forest widespread dieback of aspen groves in ranges for domestic livestock. Sci. 8:403-4 10. the parklands of Saskatchewan in the Johnston, A. M., and S. Smoliak. 1968. mid-1960’s and suggested that the The invasion of aspen into rough Reclaiming brushland in southwestern Alberta. J. Range Manage. 21:404406. dieback was triggered by the severe fescue grassland was found to be re- drought of 1961. Dead aspen trees lated to high growing season tempera- Maini, J. S. 1968. Silvics and ecology of Populus in Canada. p. 20-69. In Maini, J. were observed in some groves studied, tures, particularly during the month of June 1 and 2 years prior to establish- S., and J. H. Cayford, ed. Growth and but it is not known when mortality utilization of poplars in Canada. Dep. occurred. The 1961 drought appar- ment and to low May to September Forest. and Rural Div., Forest. Branch ently triggered a massive invasion of and annual precipitation 2 years prior Dep. Publ. 1205. to establishment. The effect of the loss aspen suckers into the grassland. The Maini, J. S. 1960. Invasion of grassland by of apical dominance in aspen on inva- in the Northern invasion was closely related to high Populus tremuloides sion into grassland by suckers was Great Plains. PhD thesis, Univ. Sas- mean temperatures and low precipita- uncertain. katchewan, Canada. 231 p. tion 1 and 2 years prior to establish- Moss, E. H., and J. A. Campbell. 1947. The ment. It is not known whether this fescue grassland of Alberta. Can. J. Res. invasion was related to the loss of Literature Cited 25 : 209-227. Shirley, H. L. 1932. Does light burning apical dominance caused by dieback of Ayyad, M. A. G., and R. L. Dix. 1964. An stimulate aspen suckers? II. J. Forest some of the mature trees. analysis of a vegetation-micro- 30:419-420.

SRljA . . . 28th Annual Meeting l Maria Isabel-Sheraton

l Mbxico City

l February IO- 14, 1975 *

*The date wasgiven erroneously as Feb. 4-9 in two pre vious issues.

266 JOURNAL OF RANGE MANAGEMENT 27(4), July 1974