By B. Patton, P. Nyren, B. Kreft, J. Caton and A. Nyren
North Dakota State University Ag. Experiment Station
Central GrasslandsResearch Extension Center, Streeter, ND
The grazing intensity study at the Central Grasslands Research Center is in its eighth year. The objectives are: to determine the optimum cattle stocking rate without damaging the rangeland resource; to develop a model to predict forage production in the spring so that livestock producers can better plan their forage requirements for the year; and to develop techniques to inventory rangeland and monitor utilization, range trend and range condition. Instruments for estimating forage production are being tested. Plant species which appear to be favored by no grazing or light grazing and species which appear to be favored by heavy grazing have been identified.
Five grazing treatments, or intensities of grazing, are included in the
study: no grazing, light, normal, heavy and extremely heavy. Normal is defined
as that grazing intensity which leaves 50% of the forage produced in an
"average" year at the end of the grazing season. The light grazing
treatment is stocked to leave 65%, the heavy treatment 35% and the extreme
treatment 20% of the forage produced in an average year. A certain amount
of trial and error is required in adjusting stocking densities, grazing
patterns and length of grazing season to achieve these grazing intensities.
Each of these treatments is applied to three pastures so that differences
due to grazing intensity can be separated from those due to natural variability
of the pastures. Changes in the vegetation are monitored on plots located
on silty and overflow range sites in each pasture. These sites are used
because they are the most common in the Coteau region. Pastures with no
grazing are simulated by fencing out areas on three silty range sites and
three overflow range sites located within the grazed pastures.
Grazing begins each year around mid-May. Table 1 gives the stocking history of the study. To keep the same level of stress on the plants each year, grazing will continue until half of the amount of forage produced in an average year remains on the pastures grazed at the normal rate. It will take several more years to determine the average productivity of these pastures. Table 2 gives peak total forage production for 1989 through 1996. Average production for 1989 to 1996 was 3,855 lbs/acre on overflow range sites and 2,798 lbs/acre on silty range sites. Therefore, an average of 1,928 lbs/acre should remain on overflow sites and 1,399 lbs/acre on silty sites at the end of the grazing season on pastures stocked at normal stocking density.
Figure 1 shows the forage remaining at the end of the grazing season for each treatment in each year of the study. Reference lines indicate the amount of forage we would like to see remaining for each grazing treatment. This shows the progress being made in adjusting stocking rates to achieve the desired use levels at the end of the grazing season.
Table 3 shows the average daily gains and gains/acre of cattle on the trial each year from 1989 to 1996. The average body condition scores for each treatment from 1994 to 1996 are also shown on table 3. This is a visual ranking of the amount of fat on an animal's body with 10 being extremely fat and 1 being extremely thin. This ranking was made because of concern that the animals on the extreme grazing treatment were coming off in poor condition. Figure 2 shows the relationship between stocking rate and average daily gain,figure 3 shows the relationship between stocking rate and gain per acre and figure 4 shows the relationship between stocking rate and economic return for 1991 to 1996. The years 1989 and 1990 are not included in these graphs because none of the pastures were stocked heavily enough to significantly reduce average daily gains. As the grazing intensity increased, average daily gain decreased. Gains per acre increase until they reach a certain level and then begin to decline, and profit per acre shows the same pattern. The stocking rate is expressed in terms of animal unit months (AUMs) per ton of forage instead of the more common AUMs/acre to make results comparable with other sites that may differ in productivity. This should also remove some of the effect of year to year variation in forage production. As is apparent from figure 2, the relationship between stocking rate and average daily gain differs significantly between years (p<0.0005). These differences may be due to variations in forage quality, the effect of weather on the animals, class of animal, their initial weight, or their potential to gain.
If cattle prices were constant, return per acre would peak at a stocking rate somewhere below maximum gain per acre with the exact point depending on carrying costs (interest, death loss, salt and mineral, vet cost, transportation, labor and land). However, fluctuating cattle prices make determining an optimum stocking rate difficult. For example, in May 1991 we stocked our pastures with bred heifers weighing an average of 800 lbs. An 800 lb heifer was valued at $702.26. Carrying cost for the season for this animal was $35.29, so if we had sold her in September when the cattle were removed from the trial we would have had to get $737.55 to break even. To get this price, assuming a 5% shrink, she would have had to come off the pasture weighing 941 lbs which would be an average daily gain of 1.18 lbs/day. Average daily gains were not very good in 1991 therefore reducing economic returns. In May 1992, we stocked our pastures with bred heifers averaging 750 lbs. Heifers of this weight were valued at $560.33 in May and at $620.33 in August at the end of the grazing season. Carrying costs for this animal in 1992 were approximately $19.77 which would have returned $40.13 even without a gain in weight. This difference in beginning and ending value of the animals made it possible for the stocking rate with maximum return/acre to exceed the stocking rate with maximum gain/acre in 1992.
Table 4 gives the stocking rates with the maximum predicted gain per acre and maximum predicted profit per acre for each year from 1991 to 1996. These values correspond to the peaks of the curves in figure 3 and figure 4. As is obvious from figure 4 and table 4 there were three years in which no stocking rate would have provided a positive return and three years with good returns. Also from table 4 you can see that in four of the six years the stocking rate that provided the greatest economic return was less than the stocking rate that resulted in the greatest gain per acre. We cannot predict what stocking rate would be optimum in any particular year and even if we could, it might not be practical for producers to be constantly adjusting their herd size to fit the current year's optimum. Included in table 4 are the gains and returns we would have expected each year with a constant stocking rate of 0.57 AUMs/ton of forage. This was the average stocking rate on our "normal" grazing treatment: the stocking rate that attempts to leave half of forage produced in an average year. In table 4 the row labeled "all years" gives the stocking rates which would have resulted in the greatest gain or profit if a constant stocking rate was used and the results averaged over the six years. Again, the stocking rate which provided the most profit was less than the stocking rate which provided the most gain. We can assume that stocking at 1.56 AUM/ton of forage would have returned an average of $20.14/acre over the years 1991 to 1996; however, losses would have been greater in poor years and returns less in good years than if one could have stocked at the optimum for each year, ranging from $20.71/acre in 1994 to $90.53/acre in 1993. We are able to determine these optimum values because we know how the cattle gained and what prices were in these years.
This trial is continuing and these relationships need more study before we can recommend the stocking rate which will give the greatest return in the long run. Stocking at the rate which produced the greatest economic return in 1992 would damage the pasture and, if prices are low, it could result in substantial financial losses. In the long run a moderate stocking rate would be most profitable. We cannot yet recommend the 1.56 AUMs/ton of forage in table 4 because it is only based on six years of data and changes in forage production or cattle prices could make it inappropriate. Also producers should keep in mind that an optimum stocking rate for a stocker operation may not be optimum for a cow-calf operation. In the future we hope to provide more information to help the livestock producer make a sound decision.
Forage production is determined by clipping plots inside of wire cages which exclude grazing, and forage remaining is determined by clipping paired plots located outside of the cages. The amount of forage used by livestock is determined by comparing the two. This sampling is very time consuming and labor intensive, but it provides the following information which can be obtained in no other way:
1. We can determine if the grazing treatments affect forage production. 1992 was the first year that forage production differed significantly among the grazing treatments and it was significant only on the silty sites. Each year since then forage production has differed among treatments on both silty and overflow range sites. Table 5 shows total forage production by grazing treatment on the silty range sites in 1992 to 1996. In 1992 as grazing intensity increased, peak forage production decreased and the heavy and extremely heavy grazing treatments produced significantly less forage than the ungrazed and the lightly grazed treatment. In 1993 at the mid-season sampling, forage production decreased with increased grazing intensity. The ungrazed and lightly grazed treatments produced more forage than the other treatments and the extremely heavy treatment produced less forage than all but the heavy treatment.
We consider peak forage production for a pasture to be the highest yield of the mid-season and end of season sampling. We usually try to time the mid-season clipping with the actual peak in forage production. 1993 was an unusually cool year, and although moisture was adequate, temperature limited plant growth until August when the rains stopped. As a result the actual peak in forage production occurred closer to the end-of-season than the mid-season sampling. By then differences in total production between treatments was not significant but grass production was still significantly less on the heavy and extremely heavy treatments than on the other treatments (Table 5).
In 1994 the normal grazing treatment produced the most forage. Production on this treatment was significantly greater than on both the ungrazed and extremely heavy grazing treatments (Table 5). This may indicate that in the longterm both nonuse and overuse can be detrimental to the productivity of native rangeland. In 1995 the extremely heavy grazing treatments produced the least forage. The other treatments were not significantly different from each other in forage production (Table 5). In 1996 the extreme treatment again produced the least forage and the light treatment produced the most (Table 5).
Production on overflow range sites only differed among treatments in end-of-season total yield in 1993 (Table 6). Here the ungrazed treatment produced significantly less forage than all but the extremely heavy grazing treatment. This was probably caused by the abundant litter on the ungrazed treatment reducing the amount of sunlight reaching the surface and limiting soil temperatures.
In 1994 overflow range sites only differed between treatments in grass production at the beginning of the season. Grass growth was 698 and 688 lbs/acre on the ungrazed and extremely heavy grazing treatments, respectively, compared to 1024 to 1096 lbs/acre on the other treatments when grazing began. This was a good year for plant growth and these treatments had caught up by the midseason sampling. In 1995 the heavy treatment produced the most forage and the ungrazed and extremely heavy grazing treatment produced significantly less forage than all the other treatments. Again in 1996 the ungrazed and extremely heavily grazed treatments produced significantly less forage than the other treatments.
2. Forage utilization data are used to help place mineral blocks and adjust livestock numbers to get the desired grazing intensity on the sample sites in each pasture. Table 7 gives the estimated percent forage disappearance at the time the cattle were removed. The term disappearance is used because forage lost due to the natural drying up and breaking off of leaves and stems is included with that consumed or trampled by cattle. Several facts are apparent from examination of this table. First, a substantial amount of the forage produced had disappeared even on the ungrazed treatment. Second, percent disappearance was less than ideal disappearance for all but the extremely heavy grazing treatment. Some adjustments will be made in the number of livestock and the location of salt and mineral blocks in some of the pastures in 1997 to bring the grazing treatments closer to the desired levels.
3. Forage production will be correlated with soil moisture and precipitation to develop a model to predict forage production. Data will have to be collected for a number of years before a model can be developed. However, we are seeing differences in available water between the different grazing treatments. On overflow range sites, lightly grazed pastures have more available water than heavily grazed pastures. The differences in available water occur during both soil water recharge and discharge. This indicates that on heavily grazed sites more water runs off during a rain and sunlight evaporates more water from the soil surface. On silty range sites, moderately grazed pastures have more available water than ungrazed or heavily grazed pastures. The ungrazed treatment has less available water because the plants on that treatment have more leaf area than the grazed plants, and more water is removed from the soil by transpiration.
4. The forage production samples are used to calibrate and test the swardstick and radiometer, two alternative methodologies for sampling forage production .
5. Forage production from known locations will be compared with reflectance values on infrared and regular color aerial photos. The photos can be scanned into a computer and analyzed to develop a map and comprehensive inventory of the entire forage base.
6. The forage samples will be analyzed each year for nutritional quality to determine if, over time, different intensities of grazing result in plant communities which produce forage of different quality. Table 8 shows the average nutritional quality of grasses and forbs on each treatment from 1989 to 1996. Although it is clear that differences in nutritional quality are developing between the grazing treatments, the reasons for the differences are not clear. On silty range sites the grasses have higher crude protein and digestibility and lower fiber components at the higher grazing intensities. On the heavily grazed treatments the grass that is available for grazing is mostly regrowth which is of higher quality. However on overflow sites both grasses and forbs are higher in fiber components at the higher grazing intensities. Perhaps on these sites cattle are selecting species of higher quality and leaving those that are higher in fiber. On silty sites forbs are lowest in neutral detergent fiber at the normal grazing intensity. As the ungrazed forage matures on the ungrazed treatment it becomes higher in fiber. On the heavily grazed treatments only forbs of lower quality would remain ungrazed. These differences in nutritional quality have occurred gradually over the course of the study.
Forage quality of the various plant species change throughout the year and can be different at the same time of year in different years. These changes are due partly to the phenological, or growth stage of the plant and partly to the climate in which the plant has been growing. Available water and soil temperatures are the major factors that affect plant growth and they may also affect forage quality.
To better understand these changes and their effect on livestock gains we began collecting samples of the seven most abundant plant species on the grazing intensity trial beginning with the 1995 growing season. Samples were collected weekly throughout the grazing season and once every two weeks throughout the dormant season, provided the ground was not covered with snow. The samples will be analyzed for crude protein, in vitro dry matter digestibility, and ash. Climatic data which includes daily temperature and precipitation is also collected.
Figure 5 shows the crude protein content and figure 6 shows the in vitro dry matter digestibility for the 1995 growing season. 1996 data is not yet available and ash is not presented because of space limitations. Although most species have their best quality in the spring, none show the same pattern and the species which had the most crude protein or highest digestibility changed through the season. One of the advantages of native rangeland is its diversity of plant species which allows livestock to select the plants which are of the highest quality at any point in time. When we have more data we will compare nutritional quality between years and try to determine what climatic factors are responsible for yearly differences in forage quality.
Changes in the plant communities are monitored by sampling the percent frequency of occurrence, density per unit area, and percent basal cover of all plant species as well as sampling the weight of herbage produced. Frequency of occurrence is sampled by placing a frame on the ground fifty times along a transect at each sampling site. Every time it is placed on the ground all the plant species which occur in the frame are recorded. The number of frames a species occurs in, divided by the total number of frames, multiplied by 100, is the percent frequency of that species. The frequency value obtained from sampling is dependent on the size of the frame. If the frame is too small, the species will rarely be recorded and if the frame is too large, the species will be present in almost every frame. Because species may differ in their abundance on a particular site, and a particular species may differ in its abundance on different sites, it is not possible to select a single frame size that is optimum for all species on all sites. For this study, a 25 x 25 cm frame with 5 x 5 cm and 10 x 10 cm frames nested within it, is being used.
Density is determined by counting all the individual plants of each species in each of the frames. Density data were only collected on shrubs and forbs from 1988 to 1991. In 1992 we began collecting density data on cespitose (bunch) grasses and sedges. Density data are not collected on rhizomatous grasses and grass-like plants because of the difficulty of counting individual plants.
Basal cover is sampled by repeatedly placing a 10 point frame on the ground and recording each time one of the points strikes the base of a plant. The number of times a point hits the base of a plant divided by the number of points read, times 100, is the percent basal cover of that species. Basal cover is stable from year to year because it doesn't change unless old plants die or new plants are established. Also the point frame can be used to sample the amount of litter or bare ground as well as total plant basal cover. However, it requires too large a sample size to provide reliable data for all but the dominant species of a plant community. The point frame was used to sample litter and bare ground in 1996. Frequency can be sampled more quickly than density or basal cover which makes it an ideal method for monitoring range trend or the direction of change in range condition. These three methods of monitoring vegetation change can be complementary and are used together in this study. Frequency data from 1988 to 1996, density data from 1988 and 1990 to 1996 and
basal cover data from 1988, 1990, 1992, 1993 and 1996 were compared for each sample site using analysis of variance. The change in abundance of species between years was determined for each site. The arcsine transformation was applied to frequency and basal cover data to convert it from a binomial distribution to a nearly normal distribution. Analysis of variance was performed to determine if there was a change in species abundance across all sites which might indicate a response to annual climatic change, or if there was a change in response to the different grazing treatments. All tests were performed at the p=0.05 level. Table 9 lists those species which appear to be favored by no grazing or light grazing and table 10 lists those species which appear to be favored by heavy grazing. In addition to the changes listed for plant species, litter has decreased on overflow range sites and bare ground has increased on both silty and overflow range sites under heavy grazing.
The value of the plant community for grazing depends on the plant species present and their forage quality. As livestock select plants of high palatability in their diet, they give a competitive advantage to plants of lower palatability and cause a shift in the composition of the plant community. An ungrazed pasture may consist of high quality forage but it is of no use to a livestock producer unless the cattle are allowed to graze it. Likewise, a pasture which has been continually overgrazed may consist of plants of low forage quality and would be of little use to a livestock producer. This research will continue to monitor the changes in the vegetation and livestock performance for four more years.