• No results found

Volume 34, Number 5 (September 1981)


Academic year: 2020

Share "Volume 34, Number 5 (September 1981)"


Loading.... (view fulltext now)

Full text


Forage Production on Important Rangeland

Habitat Types in Western Montana



Peak standing crop and its variability were evaluated over B 3-year period on 35 sites representing 13 important grassland and shrubland habitat types in western Montana. Average production ranged from slightly over 600 kg/ha on the least productive habitat type to 2,900 kg/ha on the most productive. Good sites seldom produced more than twice as much as poor sites within the same habitat type. Although PS much as 2% times more herbnge was produced on a site during the high year than during the low year, the maximum yearly dilference for all sites averaged only 1% times greater. Variation of vegetation classes and factors contributing to production differences are discussed.

The potential oi different rangeland sites ior producing forage differs greatly in the highly varied environment of the northern Rocky Mountains (Fig. I). These differences inforagcproductian. coupled with observed differences in other range values and responses to management activities, have stimulated efforts to identify and classify rangeland units according to their inherent capabilities. Much oithe recent thrust in wildland classification in the West has focused on the habitat type concept (Daubenmire

1968; Piister 1976). A classification using this concept has been developed recently for the nonforested rangelands in the moun- tainous western third of Montana (Mueggler and Stewart 1980). Once such a classification has been developed, however, resource parameters related to the classification units need to be defined. Quantification of resource values is essential to land management planning. Forage, of course, is generally accepted as the principal product to be derived from much of our western rangelands.

Forage production on native rangelands in the West is noted for its variation from year to year because of yearly weather differen- ces, and for its variation from place to place because of basic enwronmental differences. Knowledge of the average iorage- producingcapabilities oia given range typeand ofthefluctuations around this average caused by yearly weather variations is funda- mental ior planning long-term grazing capacities. The need ior information about forage potentials and variability an western Montana rangeland habitat types prompted the cooperative study between the Forest Service’s (Northern) Region 1 and the Inter- mountam Forest and Range Experiment St&on, results of which are reported here.


Of the 29 grassland and shrubland habitat types that we des- cribed ior western Montana (Mueggler and Stewart 1980), 13 of the most important were selected for evaluating peak standing crops. Importance was determined by the relative amount oiland occupied by a given habitat type and by the potential value afthe type ior producing forage ior livestock and, in some cases, ior deer and elk. The habitat types selected for measurement are listed in Table 1.

In most casts, the production level of each habitat type was evaluated by sampling 3 sites subjectively selected to represent the range in productivity for that habitat type; these sites were subse- quently categorized as poor, intermediate, or good. The Agro-

pyron spicatum/Agropyron smithii and Purshia

lridentaro/Fesluca srobrella habitat types, however, are repre- sented by only two sites each and the Deschampsia cnespitoso- /Carex spp. h.t. is represented by only one site.

The selection of stands to span the range of site productivity within a habitat type was based initially upon species canopy-cover data collected during the development of our habitat-type classifi- cation for western Montana. From the numerous stands falling within a given habitat type, one was selected to represent each site productivity category (poor. intermediate, or good) an the basis of total canopy cover and the canopy cover of individual species. Each site was free from abusive grazing and occupied by little- disturbed native vegetation. We attempted to obtain wide geogra- phical separation of the sites representingany one habitat type. In some cases, production differences subsequently measured between the intermediate and poor or intermediate and good categories were not great; however, we believe that thedifferences in production between the poor and good sites reflect fairly well the range in productivity usually encountered in a given habitat type. The locations of the 35 sampled sites scattered through western Montana are shown in Figure 2. Site numbers refer to a given habitat type in Table I.


Table 1. Habitat types sampled, site number indicated by the habitat type name.

location, Fig. l), productivity classification, and prominent on each site in addition

Habitat type Poor :

Site productivity

Intermediate Good


Stipa comatal Bouteloua gracilis

Site number: Species:

Agropyron spicatum / Bouteloua gracilis

Site number: Species:

Agropyron spicatum/Agropyron smithii

Site number: Species:

Festuca scabrella / Agropyron spicatum

Site number: Species:

Festuca scabrellal Festuca idahoensis

Site number: Species:

Festuca idahoensisl Agropyron smithii

Site number: Species:

Festuca idahoensisl Agropyron spicatum

Site number: Species:

Festuca idahoensisl Agropyron caninum

Site number: Species:

Artemisia tridentata/Agropyron spicatum

Site number: Species:

Artemisia tridentatal Festuca idahoensis

Site number: Species:

Potentilla fruticosa/Festuca scabrella

Site number: Species:


Opuntia polycantha Sitanion hystrix Astragalus purshii Sphaeralcea coccinea

9 149

Koeleria cristata Koeleria cristata Artemisia frigida Tragopogon dubius Gutierrezia sarothrae Astragalus dasyglottis Phlox hoodii Chrysopsis villosa

83 Artistida longiseta

Poa cusickii Koeleria cristata Phlox hoodii

248 Chrysopsis villosa Penstemon spp. Artemisia frigida Sphaeralcea coccinea


Festuca idahoensis Poa sandbergii Lupinus sericeus An tennaria rosea

304 Stipa occidentalis Besseya wyomingensis Antennaria rosea Pedicularis contorta


Phlox multtflora Astragalus purshii Artemisia frigida Antennaria rosea

27 Galium boreale Carex stenophylla Pedicularis contorta Koeleria cristata

125 Danthonia intermedia Carex vallicola Koeleria cristata Lupinus wyethii

8 Koeleria cristata Poa sandbergii Oxytropis riparia Carex stenophylla

272 Antennaria parvtfolia Erigeron compositus Phlox hoodii Astragalus miser

354 Festuca idahoensis Artemisia frigida Agropyron spicatum Agropyron dasystachyum


Carex filifolia Koeleria cristata Artemisia frigida Phlox hoodii

232 Festuca idahoensis Agropyron smithii Chrysopsis villosa Carex stenophylla

372 Arenaria congesta Agropyron spicatum Geum trtflorum Antennaria rosea

341 Artemisia frigida Poa cusickii Stipa comata Koeleria cristata

105 Poa sandbergii Agoseris glauca Lupinus sericeus Chrysopsis villosa


Geranium viscosissimum Po ten tilla gracilis Achilles millefolium Koeleria cristata

234 Koeleria cristata

Chrysothamnus vicidtflorus Stipa comata

Poa sandbergii


Antennaria rosea Agropyron spicatum Poa sandbergii Saxtfraga rhomboidea

365 Danthonia parryi Solidago missouriensis Agropyron spicatum Aster falcatus

358 Astragalus pectinatus Carex filtfolia Artemisia frigida Carex stenophylla

179 Artemisia frigida Phlox hoodii Koeleria cristata Poa sandbergii

87 Festuca idahoensis Lupinus sericeus Koeleria cristata Carex stenophylla

208 Antennaria rosea Artemisia ludoviciana Aster integrtfolius Lupinus sericeus

171 Astragalus cibarius Koeleria cristata Carex filfolia Poa cusickii

28 Danthonia intermedia Geum trtflorum Viola adunca Lupinus sericeus

56 Lupinus sericius Erigeron subtrinervis Galium boreale Bromus carinatus

239 Koeleria cristata Lupinus sericeus Artemisia frigida Penstemon aridus

46 Aster integrtfolius Danthonia intermedia Geum trtflorum Erigonum umbellatum

258 Galium boreale Geranium viscosissimum Festuca idahoensis Agropyron dasystachyum


Table 1. Continued

Habitat type Poor

Site productivity

Intermediate Good

Purshia tridentatal Festuca scabrella

Site number: Species:

Deschampsia caespitosal Carex spp. Site number:


246 Agropyron spicatum Bromus tectorum Balsamorhiza sagittata Poa sandbergii

371 Carex athrostachya Juncus balticus

Phleum alpinum Aster integrifolius

306 Festuca idahoensis Balsamorhiza sagittata Agropyron spicatum Bromus tectorum

‘In decreasing order of abundance as judged by canopy cover.

rosea and Phlox hoodii may have sufficient canopy cover to occur among the more abundant secondary species, they may produce considerably less biomass than the taller grasses and forbs with less canopy cover that are not listed among the four secondary species. A detailed listing of species likely to occur within each habitat type is given by Mueggler and Stewart (1980).

Standing crop on 33 of the 35 sites was measured in three consecutive years: 1974, 1975, and 1976. Production on the poor site in the Festuca idahoensisJAgropyron caninum h.t. (site 125) was measured only in 1974 and 1975, and that of the Deschampsia caespitosa/Carex spp. h.t. (site 371) was measured only in 1976. Though incomplete, data from the two sites provide productivity information not otherwise available for these important habitat types.

We attempted to schedule productivity measurements on each site at the time of peak standing crop. This meant that the earlier developing, low-elevation, and drier sites representing the Stipa comata / Bouteloua gracilis h.t. and Agropyron spicatum / Boute- loua gracilis h. t. were measured about mid-June. Later developing

sites, such as those in the Festuca scabrella/Agropyron spicatum h.t., Purshia tridentatal Festuca scabrella h.t., and Artemisia tri- dentataJAgropyron spicatum h.t., were measured from early to mid-July; and still later developing sites, such as those in the Festuca scabrella/Festuca idahoensis h.t., Festuca idahoensisJA-

Fig. 2. Thirty-five sites in western Montana where herbageproduction was measured for three consecutive years. Site numbers are referenced to a given habitat type in Table 1.


gropyron smithii h.t., and Artemisia tridentata/Festuca idahoen- sis h.t., were measured in late July or early August. The latest developing sites in the Festuca idahoensis/Agropyron caninum h.t. and Deschampsia caespitosalcarex spp. h.t. were usually measured in early or mid-August. The logistical limitations of a single two-person sampling crew accessing sites scattered over an area as large as the mountainous western third of Montana pre- vented perfect coincidence of measurement time and peak standing crop; however, errors caused by slightly early or slightly late mea- surements were probably small.

Production on each site was measured from a randomized distri- bution of 10 clusters of five plots each on a 20 X 20 m area subjectively selected to typify the site. Each plot was circular and covered 0.45 m2. The five plots in each cluster were located within easy view of one another on grid lines 1 m apart. Since one of the purposes of the study was to evaluate yearly variation in produc- tion attributable to weather, the use of permanent plots was desira- ble to reduce sampling error caused by spatial variability in the vegetation. One of the more valid techniques for assessing produc- tion is by clipping maximum aboveground standing crop. Clipping is destructive, however, affects production in subsequent years, and thus is not suited to permanent plot measurements. Therefore, we used a 1:4 ratio of clipped to estimated plots, of which the clipped plots were different each year and the estimated plots were permanent.

Production on each of the four permanent plots within a cluster was estimated as a percentage of the fifth plot, which was then clipped to ground level. All current-year growth of the clipped vegetation was sacked and later ovendried at 60°C to a constant weight for subsequent expression as kg/ha dry matter. Estimated percentages of the permanent plots were then converted to weight, thus providing weight data from five plots per cluster, or 50 plots per site. Data were recorded separately for the major plant classes: graminoids, forbs plus half-shrubs, and woody shrubs. Data were also recorded separately for the usually few major species that together comprised approximately three-fourths of the total stand- ing crop on a site. Wherever livestock used on a site prior to production measurement was a possibility, the plots to be clipped were protected by utilization cages and the estimates from the permanent plots were adjusted to herbage utilization. Production determinations from 50 plots were judged adequate to achieve a sampling precision of within &IO% of the population mean for total herbage and for total graminoid production, but only within +35% for total forbs, all at the 90% probability level (Mueggler

1976). Sampling error for forbs increased as forb production decreased and forb distribution became more erratic. This type of sampling error relates only to comparisons between sites and not to comparisons between years, because comparisons between years are based primarily on permanent plot records.

To facilitate use of this production information by range manag- ers, we partitioned the data into forage and nonforage categories.


Table 2. Peak standing crop (dry kg/ha for important rangeland habitat types in western Montana averaged over a 3-year period with the limits of vari- ation expected in 8 out of 10 years.


Site productivity

Intermediate Good

Habitat type

Average Yearly Percent Average Yearly Percent Average Yearly Percent production variability forage production variability forage production variability forage

Stipa comataj Bouteloua gracilis

Graminoids Forbs Shrubs


Agropyron spicatum/ Bouteloua gracilis

Graminoids Forbs Shrubs


Agropyron spicatum / Agropyron smithii

Graminoids Forbs Shrubs


Festuca scabrellal Agropyron spicatum

Graminoids Forbs Shrubs


Festuca scabrellal Festuca idahoensis

Graminoids Forbs Shrubs


Festuca idahoensisj Agropyron smithii

Graminoids Forbs Shrubs


Festuca idahoensisl Agropyron spicatum

Graminoids Forbs Shrubs


Festuca idahoensisl Agropyron caninum

Graminoids Forbs Shrubs


Deschampsia caespitosal Carex spp.

Graminoids Forbs Shrubs Total 263 7 0 270 +I26 It 9


100 351

0 436

- 0

98 787

+I18 f117



f234 55

783 166 0 949

f188 f 95


100 14 - 85

272 * 50 213 f 53

0 -

485 +100

100 0


240 +109

248 + 30

0 -

488 f82

100 523 &IO0 100

32 343 +177 17

- 0 - -

65 866 f269 67

538 5134 100 832 +226

214 +114 0 67 +44

2 + 3 0 0 -

754 5201 72 899 +245


20 -


- -

757 +104

142 +129

0 -

899 + 85

100 I5 87 888 109 0 997

&I98 100 888 f409

+ 22 13 442 +194

- - 17 + 8

+177 91 1347 f610

100 0 0


934 f 10 100

268 +221 2

0 - -

1202 f231 75

1097 500 0 1597

f 86 5105

- f 59

100 8


1692 +655 100

136 f 36 0

0 - -

1828 z!I 745 93

397 + 50 100 985

386 k228 22 21

24 f 19 0 0

807 5229 60 1006

+3wl 100 1253

+ 12 0 195

- - 0

+397 98 1448

f128 It150 f227 100 50 - 93 304 432 0 736

f 45



412 _I

925 -

0 -

1337 -

100 20 - 53 .lOO 14 41 - - 100 13 0 40 100 13 1 34

583 + 18 100 521

247 f 98 30 927

4 f 4 100 0

834 +112 79 1448

+193 100

f226 14

f367 45

750 976 0 1726 +240 5492 +727

96 1017

33 847

- 0

61 1864

+203 k186 +288 100 14 - 60 - - - - - - 2877 33 0 2910 -2 - - - 100 0 - 99 - - - -

Artemisia tridentataj Agropyron spicatum

Graminoids Forbs Shrubs


Artemisia tridentatal Festuca idahoensis

Graminoids Forbs Shrubs


312 It101

149 f 35

374 f 67

835 f 145

411 f142 100 479 f 84 100

57 + 30 0 259 + 22 0

374 f 74 1 238 f 29 0

842 f 56 49 976 f 80 49

5225 f 26

474 f 177 156 + 57 855 f219

259 f 16 100 683 f 71 100

457 f 65 6 774 f112 11

492 + 85 0 161 f 57 0

1208 f104 24 1618 f 93 48


Table 2. Continued

Habitat type

Potentilla fruticosa/ Festuca scabrella

Graminoids Forbs Shrubs Total

Purshia tridentata / Festuca scabrella

Graminoids Forbs Shrubs


Site productivity

Poor Intermediate Good

Average Yearly Percent Average Yearly Percent Average Yearly Percent production variability forage production variability forage production variability forage


758 +206 100 929 +314 100 I295 +291 too

217 f123 12 312 +I24 2 502 +194 31

56 f 27 0 48 f 49 0 189 f134 22

1031 f333 76 1289 f464 73 1986 +574 75

456 f242 100 483 i 86 100

425 k188 88 607 f308 81

123 fl09 100 298 5139 99

1004 +529 95 1388 +369 91

‘Only 2 years of data available. ?Only I year of data available.

A species was considered a forage contributor if it rated fair or better in palatability to sheep and/or cattle. The palatability rat- ings were obtained from Mueggler and Stewart (1980). Weight data on a site for those species that were not measured separately (approximately 25% of the biomass) were inferred from the species’ relative canopy cover that was available from a previous study.

tion of forage plants indicated that the most productive habitat type was seven times more valuable as livestock range than the least productive habitat type:

Results and Discussion

Variation between Habitat Types

A IO-fold difference in annual aboveground standing crop was encountered among these 35 rangeland sites in western Montana. Production ranged from 270 kg/ha on the poor Stipa comata/ Bouteloua gracilis site to 2,9 10 kg/ ha on the Deschampsia caespi- tosal Carex spp. site (Table 2). Practically all vegetation on these two extreme sites consisted of forage species.

The habitat types were ranked according to overall productivity by averaging the data for the different sites within habitat types over the 3-year sampling period. In most cases, the resulting pro- duction figure for a habitat type represents a nine-sample mean. This ranking showed that the most productive habitat type aver- aged almost five times more total herbage than the least productive habitat type:

Habitat type

Total standing crop W/ha)

Agropyron spicatum/ Bouteloua gracilis 613 Stipa comatal Bouteloua gracilis 669 Agropyron spicatum/Agropyron smithii 832 Artemisia tridentataf Agropyron spicatum 884 Festuca idahoensisl Agropyron spicatum 1006 Festuca scabrellal Agropyron spicatum IO82

Festuca idahoensisl Agropyron smithii 1087 Purshia tridentatal Festuca scabrella II97 Artemisia tridentatal Festuca idahoensis 1227 Potentilla fruticosaf Festuca scabrella 1436 Festuca scabrellal Festuca idahoensis 1562 Festuca idahoensisl Agropyron caninum 1646 Deschampsia caespitosa/Carex spp. 2910

Total standing crop in these natural communities does not coin- cide necessarily with their value for producing livestock forage. The relative amount of standing crop consisting of forage species ranged from about 40% of the Artemisia tridentata-dominated sites to 99% on the Deschampsia caespitosal Carex spp. h.t. (Table 2). Usually the shrub sites (excluding those dominated by Purshia tridentata) and those herblands with high proportions of forbs had the greatest percentage of production consisting of nonforage plants. A ranking of the habitat types according to total produc-

Standing Crop of forage species

Habitat type W/ha)

Agropqv-on spicatuml Bouteloua gracilis 390 Artemisia tridentatal Agropyron spicatum 408 Artemisia tridentatal Festuca idahoensis 453 Stipa comatal Bouteloua gracilis 502 Festuca idahoensisl Agropyron spicatum 567 Agropyron spicatum/Agropyron smithii 698 Festuca scabrella/Agropyron spicatum 860 Festuca idahoensisl Agropyron caninum 909 Festuca idahoensisl Agrop_vron smithii 939 Potentilla.fruticosa/ Festuca scabrella 1072 Purshia tridentatal Festuca scabrella I II0

Festuca scabrellaj Festuca idahoensis 1260 Deschampsia caespitosalcarex spp. 2880

These herbage production figures, however, are not equivalent to the amount of feed available for use by livestock. The figures must be discounted to what is considered proper use, which is contingent primarily upon plant requirements for maintaining vigor. Requirements vary for different forage species and season of use. Effects of use are reviewed in general by Ellison (1960) and Jameson (1963), and as related directly to these habitat types by Mueggler and Stewart ( 1980). Seldom is more than 50% use of key species considered wise under season-long grazing. A greater pro- portion of the total production of forage species might be removed without long-term ill effects under fall grazing or under such a specialized grazing system as rest-rotation.

Munn et al. (1978) examined eight of these habitat types to determine environmental factors that might be related to differen- ces in productivity. They identified 18 variables significantly related to total standing crop and concluded that soil morphologi- cal characteristics are more useful predictors of productivity than either estimated climatic data or soil nutrient data. The thickness of the mollic epipedon (the relatively thick, dark surface horizon of mineral soil) was most highly correlated (rx0.73) with productiv- ity. They indicated that productivity generally increased along an increasing moisture gradient and found significant correlation (r=O. 59) betw een estimated annual precipitation and total production.

We examined the relationship between estimated average pre- cipitation and standing crop on 12 of the 13 habitat types for which we had production data. The Deschampsia caespitosa / Carex spp. h.t. was omitted because it typically occurs in mountain meadows subject to spring flooding or subirrigation. The long-term precipi-


tation average for each site was first estimated from isohyetal maps of western Montana (Ross and Hunter 1976). Average precipita- tion on each site for the 1974-1976 period was then determined by adjusting the isohyetal figure according to the yearly departure from normal during this period of the permanent weather station (U.S. Weather Bureau, 1973-1976) closest to each site. The adjusted precipitation for the three sites within a habitat type was then averaged to obtain a mean precipitation value for the habitat type for the 1974-1976 study period. We found that the average production for each of the 12 habitat types was significantly corre- lated (rx0.73) with the average October through September pre- cipitation for the 3-year period.

these relatively undisturbed sites within habitat types was only slightly greater than the one and one-half to two time difference found in total vegetation. We would expect, however, the propor- tion of forage plants on a site to decrease appreciably under abusive grazing. This would probably result in substantially greater within-type variation of forage production than of total production when various range condition classes are involved.

Variation within Habitat Types

Individual plant species are known to exhibit tolerance to cer- tain variations in environmental conditions. Some species, e.g., certain Astragalus, are rather site specific. Others, e.g. Festuca idahoensis, can grow under a fairly wide range of environmental conditions; in other words, they have broad ecological amplitude. A species’ ecological amplitude is governed by phenotypic plastic- ity and by the genetic diversity within the species’ classification unit, which in turn is sometimes dependent upon rather arbitrary classification criteria. The same phenomenon occurs with plant communities or, in this case, habitat type classification units. The environmental specificity of a habitat type is governed in part by the ecological amplitude of its defining species; it is also strongly influenced by the amount of diversity in secondary species arbitrar- ily permitted in development of the vegetation classification unit. Thus a given habitat type can occupy a restricted range of environ- mental conditions.

We believe that the amount of soil water available for plant growth is the primary factor affecting both within-type production differences and production differences between habitat types. Available soil water is a function of precipitation, soil depth for water storage, and those factors affecting water loss by evapotrans- piration. We were unable, however, to demonstrate statistically a relationship between production within habitat types and either precipitation or soil depth because of inadequate data for each t Y Pee

Variation between Years

We obtained an estimate of the influence of yearly weather differences on herbage production by sampling each site over a 3-year period. From the amount of variation in production encountered during this period, we computed the production limits likely to occur in 8 out of 10 years-the yearly variability data presented in Table 2.

By purposefully selecting what we intended would represent poor and good site conditions within a given habitat type, we obtained an approximation of the variability in herbage produc- tion likely to be encountered within the environmental range of the type. The poor-site versus good-site differences in production shown in Table 2 are 3-year means, thus minimizing production differences between sites that might be caused by yearly weather variations.

Over this particular 3-year period total standing crop on one site during the high year was as much as 2% times greater than during the low year, but over all sites averaged only about l-1/2 times greater. The coefficient of variation in total standing crop ranged from about 5 to 40% and averaged 22%. There was no pronounced difference between either habitat types nor between good or poor sites in the proportionate amount of yearly variability in total standing crop.

Less than twice as much total standing crop was produced on the good than on the poor sites in all but one habitat type. The Stipa comata/Bouteloua gracilis h.t., one of the least productive grass- lands in western Montana, produced over three times as much vegetation on the good than on the poor site. Within habitat type variability in production by vegetation class was substantially greater than that for total vegetation. Generally from two to three times as many graminoids and frequently over three times as many forbs were produced on the good site as on the poor sites. Differen- ces in production of individual species often were even greater. Thus, total production was generally least variable, graminoid production next, and forb production most variable between sites within a habitat type.

Standing crop of total graminoids varied about the same amount from year to year as total vegetation. The greatest varia- tion was a difference of about 2’55 times on one site. Graminoid production on all sites averaged 1.6 times greater during a high production year than during a low production year. The range and average coefficient of variation for yearly variability of graminoid production was essentially the same as that for total standing crop. No relationship was apparent between the relative amount of variability and either habitat type or site quality.

The standing crop of total forbs varied about twice as much from year to year as the graminoids. The amount of variation appeared to be related to the quantity of forbs in the community. Forb production over all sites averaged 2.6 times more during a high production year than during a low production year. Where forbs made up less than 20% of the total biomass, production increased an average 3.7 times compared to an average 2.0 times where forbs made up more than 20% of the biomass. The coefficient of varia- tion for the former sites ranged from approximately 50 to 100% and for the latter sites, from 10 to 50%.

We found that the difference in amount of forage plants between Yearly variation in annual shrub production was similar to that

Table 3. Comparison of peak standing crop (dry kg/ha) between a 3-year and a lo-year period on two sites in theFestucu idahoensis/Agropyron spicatum

habitat type.

Ten-year record (1964-l 973) Three-year record ( 1974- 1976)

Range in Average Yearly


Range in Average production production variability’

Yearly production production variability’

Site 27

Graminoids 270-586 407 f 44 256-332 304 + 45

Forbs 276-42 1 359 k 22 3 16-502 432 +110

Total 661-917 766 + 41 572-825 736 +155

Site 28

Graminoids 228-628 446 f 49 352-706 521 It193

Forbs 583-1451 1024 f 94 708- 1123 927 +226

Total 81 I-1774 1470 +119 1061-1657 1448 f367

‘Limits of variation expected in 8 out of IO years.


of the forbs. When all sites were averaged, current annual produc- tion of shrubs was 2.8 times greater during a high production year as during a low production year. Where shrubs made up less than 20% of the total biomass, four times more was produced during the high than during the low year. Where shrubs constituted more than 20% of the total biomass, shrub production was only 1.6 times greater on the average. The coefficient of variation in early shrub production ranged from about 30 to 90% for those sites with less than 20% shrubs and from about 10 to 4OY0 for those sites with more than 20% shrubs.

How representative of long-term variability is the 3-year period covered by this study? October through September precipitation records from U.S. Weather Bureau stations closest to the sites indicate that the 1974 through 1976 period average 15% wetter than the long-term averages. The wettest year of the 3-year period averaged approximately 3570 more precipitation and the driest year approximately 15% less than the long-term averages.

How typical the 3-year standing crop data are of a longer period of time is indicated by a comparison with similar data obtained over 10 years on site 27 and 28 in the Festuca idahoensis/Agro- pyron spicatum h.t. (Table 3). The IO-year records were collected

from 1964 to 1973 near the plots used for the 3-year records. Sampling intensity and methods were similar for both periods. Part of the difference in average production between the two periods can be attributed to the difference in placement of the sample plots. Therefore, variability contrasts are more easily understood when expressed as proportions of mean production.

The proportionate range in production of total vegetation on site 27, a poor site, was approximately the same for the short- and long-term records. The difference in total production between the high and low years amounted to approximately one-third of the production mean. Although the proportionate range in production of total forbs was also approximately the same for the two periods, the 3-year period poorly represented the long-term range in pro- duction of total graminoids. On site 28, which is considered a

highly productive site in this habitat type, the proportionate range in production of total vegetation from the long-term record was greater (66% of the mean) than from the short-term record (4 1% of the mean). Similar differences in proportionate ranges occurred for the two vegetation classes. This suggests that although the 3-year records may be fairly representative of the long-term varia- bility in total standing crop on poor sites, they may be conservative in portraying long-term variability on good sites. The computed ranges in variability for 8 out of 10 years shown in Table 3 are appreciably greater for the 3-year than for the IO-year record because of the difference in number of sample years.

Literature Cited

Daubenmire, R. 1968. Plant Communities, a Textbook of Plant Synecol- ogy. Harper and Row, New York, 300 p.

Ellison, L. 1960. Influence of grazing on plant succession of rangelands. Bot. Rev. 26:l-78.

Jameson, D.A. 1963. Response of individual plants to harvesting. Bot. Rev. 29:532-594.

Mueggler, W.F. 1976. Number of plots required for measuring productivity on mountain grasslands in Montana. USDA Forest Serv. Res. Note IN-T-207, 6 p.

Mueggler, W.F., and W.L. Stewart. 1980. Grassland and shrubland habitat types of western Montana. U.S. Dep. Agr. Forest Service Gen. Tech. Rep. INT-66, 154 p.

Munn, L.C., GA. Nielsen, and W.F. Mueggler. 1978. Relationship of soils to mountain and foothill range habitat types and production in western Montana. Soil Sci. Sot. Amer. J. 42: 135-139.

Pfister, R.D. 1976. Land capability assessment by habitat types. Proc. 1975 Natl. Conserv. Sot. Amer. Forest 1976: 312-325.

Ross, R.L., and H.E. Hunter. 1976. Climax vegetation of Montana based on soils and climate. U.S. Dep. Agr. Soil Conserv. Service, Bozeman, Mont. 64 p. (processed).

U.S. Weather Bureau. 1973-1976. Climatological data: Montana section. U.S. Dept. of Commerce, NOAA, Washington, D.C.


Tiller Length vs. Tiller Weight: Applications

to Plant Growth Studies



The growth and development of individual grass tillers is fre- quently monitored in intensive studies of plant response to treat- ments or environmental parameters. Since growth may be recognized as an increase in tiller weight or tiller length over time, knowledge of the relationship between them is important if researchers are to select the measure most appropriate for their needs. To make these comparisons, the mean weight and length of tillers from perennial ryegrass (Loliumperenne L.), colonial bent- grass (Agrostis tenuis Sibth.), softchess (Bromus mollis L.), and annual fescue (Vulpiu Spp.) were measured on each of seven dates during the spring growing period, 1977. Sample sizes required to estimate mean tiller length were calculated on each date using Stein’s two-stage procedure. The relationship between mean tiller weight and mean tiller length over time was evaluated by least squares analysis of a linear regression model. Results of these analyses indicated that mean tiller length can be estimated satisfac- torily from a substantially smaller sample size than can mean tiller weight. In addition, a strong linear relationship was observed between mean tiller weight and mean tiller length over time. The strength and linear nature of this relationship suggests that both measures will yield similar relative growth curves when sequential observations are plotted over time. Actual growth curves plotted for perennial ryegrass and colonial bentgrass support this supposi- tion. Therefore, since tiller lengths require a small sample size to estimate each mean, both time and money can be saved by basing growth analysis on observations of tiller length rather than on tiller weight.

Plant growth is often monitored to assess the effects of treat- ments on plant productivity and development in range and pasture studies. In studies of inter-specific plant competition and plant- herbivore interactions, the distribution of plant growth during the growing season may be as important as the total amount of aerial biomass accumulated over the entire season. The study reported here grew out of a need to determine the relative growth curves of different grass species growing together in a mixed sward. High variability in plant distribution and plant size within the sward, the relatively small area available for study, and the need to measure growth for each species separately made the use of plot-based or plant-based techniques impractical. The intensive nature of the study suggested the use of a tiller-based technique.

In intensive growth analyses, individual grass tillers often serve as the unit of observation for determining plant growth. Plant growth is generally measured quantitatively as an increase in weight or height over time. For example, Nelson et al. (1978) described growth rates in several populations of tall fescue (Festuc arundinacea, Schreb.) in mg/day and Beaty et al. (1978) reported Author is assistant professor, Rangeland Resources Program, Oregon State Uni- versity, Corvallis 97331.

This article is submitted as Technical Paper No. 4999. Oregon Agricultural Experi- ment Station, Corvallis, Oregon 9733 I.

Manuscript received January 7, 1980.

tiller weights in their study of tall fescue development while McCarty (1938), McCarty and Price (1942) and Joy et al. (1954) defined plant growth as an increase in height of individual plants. Since growth may be expressed in either units of weight or height, it is important to know the relative efficiency of using each unit and the relationship between them. The purpose of this paper is to evaluate (1) the relative efficiency of using tiller weight vs. tiller length as a measure of growth in perennial ryegrass and colonial bentgrass and (2) the strength of relationship between mean tiller length and mean tiller weight over time for perennial ryegrass (Lolium perenne L.), colonial bentgrass (Agrostis tenuis Sibth.), annual fescues (Vulpia spp.), and softchess (Bromus mollis L.).

Materials and Methods

This study was conducted on two contiguous plots, one fertilized with 34 kg/ ha of ammonium nitrate on May 1, 1977, and the other an unfertilized control plot. The study area was located approxi- mately 1 km west of Corvallis, Oregon. Elevation is approximately 100 m. Annual precipitation is approximately 100 cm. Both plots supported a mixed stand of vegetation containing approximately 40% subterranean clover (Trfolium subterraneum L.), 26% peren- nial ryegrass, 18% colonial bentgrass, 5% annual grasses, and I 1% broadleaf forbs by cover on March 30, 1977. These two plots were repeatedly sampled on seven dates during the spring growing season (between May 5 and June 10) in 1977. Care was taken to disturb the plots as little as possible during sampling. Traffic on the plots was not observed to be a problem. For each sampling date, the first 25 tillers each of perennial ryegrass and colonial bentgrass together with the first 25 tillers of annual grass encountered by the pins of randomly placed ten-point frame (Sharrow and Tober 1979) were sampled. This sampling regime was designed to select both vegetative and reproductive tillers in proportion to their density in the population. In the initial sampling (May 5), all tillers were vegetative. As plant phenology progressed through the grow- ing season, the ratio of vegetative/ reproductive tillers in the sample decreased. Virtually all reproductive tillers were encountered on May 5, May 5, May 27, and June 10 for annual fescues, perennials ryegrass, softchess, and colonial bentgrass, respectively. Sample tillers were clipped off at the soil surface, then taken to the labora- tory where they were oven-dried at 500 C for 72 hours. Tiller length was then determined to the nearest millimeter by straightening the tiller along a meter stick. Length was measured from the base of the culm to the tip of the longest leaf. Oven-dry tillers were subse- quently weighted to the nearest hundredth of a milligram on an analytical balance.

The least squares relationships between mean tiller length (X) and mean tiller weight (Y) were determined using a simple linear regression procedure (Steel and Torrie 1960). The resulting regres- sion equations of fertilized vs. unfertilized plots were compared for homogeneity of regression coefficients. No differences were evi-


dent. Therefore, data from fertilized and unfertilized plots were pooled for all further analysis.

Table 2. Coefficient of variation (C.V.) and minimum sample size (n) required to estimate mean perennial ryegrass tiller length and tiller weight on seven dates.

Results and Discussion

The high r2 values observed for all grass species examined in this study indicate that the linear relationship between mean tiller lenght and tiller weight over time was strong (Table 1). Mean tiller length is a powerful predictor accounting for 65 to 94% of the observed variability in mean tiller weight. Standard errors of the estimate were moderately low, ranging from 9 to 28% of mean tiller weight. The lowest r2 (0.65) and highest Sy.x expressed as a percen- tage of the mean (28%) occurred in the annual fescue group. Table 1. Least squares estimates for simple linear regression between mean

tiller length (mm) = X and mean tiller weight (mg) = Y of four grass species based upon n observations.

Plant species Regression line n r2 G.x (mg)

Perennial ryegrass Y = 0.60X - 40.6 14 0.94* 30.0 Colonial bentgrass Y = 0.31X - 6.9 14 0.94 8.3 Annual fescue Y = 0.13X - 8.8 9 0.65* 10.2 Softchess Y = 0.24X - 20.3 10 0.94* 5.7

‘Significant at KO.01.

Annual grasses can be difficult to identify in the vegetative stage even with the assistance of a vegetative key. While identification was not a problem with a distinctive species, such as softchess, it was a problem with relatively nondescript species such as the annual fescues. Even though all of the plants included in this group had similar tiller morphology, the inclusion of several species into a group, either intentionally or through misidentification, substan- tially weakened the length/ weight relationship. It thus appears the grouping of species for the purpose of estimating a common mean tiller length/weight relationship should be avoided even if the plants have similar morphologies.

The strong linear relationship between mean tiller length and mean tiller weight indicates that these two units of measure will yield similar growth accumulation curves for the grasses studied with the possible exception of the annual fescues (Figs. I and 2). The growth accumulation curves based upon mean tiller weight and mean tiller length plotted in Figures 1 and 2, when compared for homogeneity of regression coefficients, revealed no meaningful differences (p>. 10). Therefore, proportional rates of growth calcu- lated from either length increase or weight increase over time should be similar when expressed as a percentage of total growth observed. This concept is further substantiated by Scott’s (1961) work in which increase in leaf length and biomass gave similar proportional growth rates for three perennial tussock grasses.


1’0 20 I 30 I 40 I 50 I S’O

Days Since 5 April 1977


C.V. (%)

Length Weight

nI Length Weight

OS/OS/ 77 27 46 29 85

26104177 32 95 41 364

05/05/ 77 36 107 52 462

13/05/77 26 63 27 159

20/05/ 77 21 56 17 124

27105177 16 46 I1 84

10/06/ 77 24 51 22 106

‘Sample size required to estimate the mean k 10% with 950/0confidence. (Stein 1945)

Although similar estimates of proportional growth rate may be obtained from either tiller length or tiller weight, the efficiency of obtaining these estimates varies substantially for the two measures. The coefficient of variation was generally two to three times as great for samples of tiller weight as it was for samples of tiller length in both perennial ryegrass (Table 2) and colonial bentgrass (Table 3). As a result of the higher variation observed in tiller weights, more tillers must be examined to estimate mean tiller weight than are required to estimate mean tiller length at a given level of accuracy and confidence. For example, on April 5 a sample containing 85 ryegrass tillers was required to estimate mean tiller weight within 10% of the true mean with 95% confidence. Only 29 tillers are required to secure an equivalent estimate of tiller length. Observations indicate that the amount of time required to locate and measure an individual tiller was approximately the same regardless of whether length or weight was measured. Therefore, the smaller sample size required to estimate mean tiller length

Table 3. Coefficient of variation (C.V.) and minimum sample size (n) required to estimate mean colonial bentgrass tiller length and tiller weight on seven dates.


C.V. (%) Length Weight


Length Weight

05/04/77 20 46 16 83

26104177 23 91 21 331

05/05/77 29 75 34 225

13/05/77 36 66 52 164

20/05/ 77 18 60 13 144

27105177 19 59 14 139

10106177 26 52 27 108

‘Sample size required to estimate the mean f IOYP with 95% confidence. (Stein 1945).

loo- ATIllER LENSTII --- .“& :, .J 0 Tllltl WEIIHT . . . 0 : I’ ?I


..,’ : ,’ 4 A 0, : ,’


ISO- .,

* f 0 0.. . . A* r*o l .* A

:! 20-



,>>L.;;:_5:Z--- .“A’--

0 I I I I I I I

10 20 30 40 50 60

I 70 80

Days Since 5 April 1977

Fig. 1. Relative growth curves of perennial ryegrass tillers based upon mean tiller length and mean tiller weight.

Fig. 2. Relative growth curves of colonial bentgrass tillers based upon mean tiller length and mean tiller weight.


compared to mean tiller weight represented a true increase in efficiency directly proportional to the ratio of the sample sizes required. this increase in efficiency ranged from approximately 300 to 900% and 300 to 1600?& for perennial ryegrass and colonial bentgrass, respectively. Thus, growth analyses based upon obser- vations of tiller lengths may be done in one third or less of the time requires to perform a similar analysis based upon tiller weights.

The harvest of tillers which is required to measure tiller weight is a further disadvantage of that technique. Tiller removal not only reduces the population of tillers available for future study but may also effect plant growth by reducing the plant’s photosynthetic area. Therefore, tiller length, a non-destructive measure, may be preferable to tiller weight where objectives of the experiment or small plot size require that individual plants be sampled several times during the growing season.


A strong linear relationship was observed between length and weight of tillers from three of the four species examined. The relatively high r2 and low Sy.x values obtained for these three species indicate that tiller length is a sensitive predictor of mean tiller weight. More importantly, however, the strength and linear nature of this relationship suggests that plant growth curves derived from either mean tiller length or mean tiller weights mea-

surements will be similar. Since tiller lengths require a smaller sample size to estimate the mean, both time and money can be saved by basing growth

rather than tiller weight.

analysis on observations of tiller length

Literature Cited

Beaty, E.R., J.W. Dobson, and A.E. Smith. 1978. Tall fescue tiller weights, green forage present, and forage IVDMD. Agron. J. 70:223-226. Joy, C.R., L. Helwig, T. Reiger, and M. Supola. 1954. A comparison of

grass growth on different horizons of three grassland soils. J. Range Manage. 7:212-214.

McCarty, E.C. 1938. The relation of growth to varying carbohydrate content in mountain brome. U.S. Dep. Agr. Tech. Bull. 598. 24 p. McCarty, E.C., and R. Price. 1942. Growth and carbohydrate content of

important mountain forage plants in central Utah as affected by clipping and grazing. U.S. Dep. Agr. Tech. Bull. 818. 51 p.

Nelson, C.J., K.J. Treharne, and J.P. Cooper. 1978. The influence of temperature on leaf growth of diverse population of tall fescue. Crop Sci.


Scott, D. 1961. Methods of measuring growth in short tussocks. New Zealand J. Agron. Res. 4:282-285.

Sharrow, S.H., and D.A. Tober. 1979. A simple, lightweight point frame. J. Range Manage. 32:75-76.

Steel, R.G.D., and J.H. Torrie. 1960. Principles and Procedures of Statis- tics. McGraw-Hill Book Co., New York. 481 p.

Stein, C. 1945. A two-sample test for linear hypothesis whose power is independent of the variance. Ann. Math. Stat. 16:243-258.


Impact of Burning Pinyon-Juniper Debris on

Select Soil Properties


Burning had the greatest impact on soils beneath burned debris piles. Electrical conductivity, phosphorus, potassium, percent nit- rogen, and percent organic carbon increased significantly at all soil depths the first year after burning debris piles. No impact was evident on phosphorus, percent nitrogen, and percent organic carbon by the second year. Impacts on burned interspace areas were generally less pronounced and few impacts were measured the second year. Impact of burning on soil pH was minor.

Burning as a management tool is becoming increasingly impor- tant within the pinyon-juniper (Pinus spp.-Juniperus spp.) type. Wright et al. (1979) have recently reviewed the role and use of fire in pinyon-juniper plant communities. Wells et al. (1979) have indicated that under rangeland - conditions, effects of fire on organ- isms and soil are highly variable. From the soils standpoint,-fire may significantly affect available nutrient supplies, at least initially. The amount of nutrients released either in the form of volatiles or ash depends on the quantity available in the above- ground vegetation consumed by the fire. Studies within various vegetation types seem to indicate that the nitrogen status of soils following fire varies greatly while burning does not appreciably reduce the amount of phosphorus, potassium, calcium, magne- sium, and various micronutrients (Mueggler 1976). Wells et al. (1979) indicate that conflicting changes in chemical properties of range soils (particularly nitrogen content and phosphorus and potassium levels) appear to result from widely varying fuel charac- teristics and loading and fire intensity. Buckhouse and Gifford (l976b) found significant increase in potassium and phosphorus in overland flow following a control burn on a pinyon-juniper debris- in-place site in southeastern Utah. No significant treatment changes were detected for sodium, calcium, or nitrate-nitrogen.

The objective of this study was to determine the impact of burning pinyon-juniper debris on a site in southeastern Utah on pH, conductivity, phosphorus, potassium, percent nitrogen, and percent organic carbon of 2.5 cm incremental soil surface depths to a total depth of 10 cm.

Site Description and Methods

The study site is located near Coyote Flat, approximately 1 km west of Utah Highway 621, between Natural Bridges National Monument and Mexican Hat, Utah, and approximately 70 km west of Blanding, Utah. The study area is located at an elevation of 2,150 m and is within the confines of the Colorado Plateau.

Soils at the study site are wind derived (Aridic Arguistolls-Typic Argiustolls association) from a sandstone parent material, and may extend to a depth of approximately 1.5 m. The pH of the soil is slightly basic, averaging about 8.0. Organic matter content is low,

Author is professor and chairman, Watershed Science Unit, College of Natural Resources, Utah State University, Logan, 84322.

This study was part of a cooperative project between the Utah Agricultural Experi- ment Station (Project 749) and the Bureau of Land Management. Technical Paper Number 2477, Utah Agricultural Experiment Station, Utah State University, Logan, 84322.

Manuscript received November 19, 1979.

slightly less than 2.Oa/,. Soil texture is a sandy loam, with few rocks present.

The pinyon and juniper trees were double-chained with debris- in-place in 1967. The site (about 14 ha) was then broadcast seeded at 9.1 kg/ ha to crested wheatgrass (Agropyron cristatum) and fenced to exclude livestock.

The area was burned on September 5, 1974. At the time of the burn air temperature was 29” C, relative humidity was 10 to 12%, and the wind was somewhat variable at 17 to 25 km/ hr. The crested wheatgrass fuel load (current year’s growth plus accumulated lit- ter) was about 845 kg/ ha.

Temperature extremes within debris piles and in the interspace areas were determined using Tempils with a melt range from 52” C to 927” C. Forty-six replicated measurements were taken at a 2.5 cm soil depth and at a height of 10 cm above the soil surface. Table

1 shows mean temperatures encountered during the burn. Table 1. Mean temperatures, debris-in-place controlled burn (as defined

by the use of Tempils with defined melting points from 52” C to 927” C).

Open, grassy

interspace area Debris pile

10 cm above soil surface 187°C >777O c

2.5 cm soil depth <55” c 288’C

Total energy expended at 46 locations within the burned area was determined using water-can fire analogs as described by George (1969). These analogs were gallon and quart unused paint cans that were painted a flat black, filled with a known weight of water, and placed at ground level. Holes about 1.5 cm in diameter were punched in the cans immediately prior to the burn and then the holes were taped shut following the burn. Energy expended at each sampling point was simply approximated by determining the weight of water lost from the cans in grams and multiplying by 540 calories, the latent heat of vaporization of water at 100°C. Since water temperature in the cans at the time of the fire was approxi- mately 27” C, the energy calculations are obviously on the conser- vative side (Table 2).

Table 2. Approximate debris-in-place site.

energy release at sampling points on burned

Sampling position Interspace (grassy) area Live juniper tree (1.4 m tall) Debris pile (I tree) Debris pile (2 trees) Debris pile (3 trees) Edge of debris pile (3 trees) Debris pile (4 trees) Debris pile p6 trees) Ephedra shrub

Approximate energy release (calories X 103)

10.2 to 65.9 120.4 to 1,064.3 305.6 to 1,403.5 lJ59.5 to >1,591.4 1,070.3 to >l,601.6

37.5 to 515.7 661.5 to >1,600.0 1,126.9 to >1,598.4

14.6 to 167.4


Fig. 1. Mean average. monthly rainfall distribution in Nsukka based on IO years
Table 1: Mean dry wt, number/grazing day, area/patch, consistency of dung, and proportion of paddock fouled at each grazing during 2 seasons
Fig. 2. Number of lotebushplants that emergedonfieldplots in average (maximum and minimumj2) daytime temperature
Table 1. Description, costs and prices for the major activities in the model.


Related documents