LONG-TERM IMPORTANCE OF GRAZING, FIRE AND WEATHER PATTERNS ON EDWARDS PLATEAU VEGETATION CHANGE

Samuel D. Fuhlendorf and Fred E. Smeins

Rangelands of Texas and the world have undergone many changes over the past 100-150 years (Archer 1994). These changes have been associated with several variables, including grazing, weather patterns and altered fire regimes and for the Edwards Plateau they have caused a major decline in livestock carrying capacities (Smeins et al., this volume). It is difficult with traditional, short-term studies to isolate the importance of each variable independently. Long-term data from the Sonora Research Station provides the opportunity to evaluate these changes and predict the influence of certain management practices on the sustainability of livestock production.

The dominant change on most rangelands is the increase in abundance of woody plants. On rangelands that were previously fairly open, an increase in woody plant abundance can cause a dramatic reduction in the livestock carrying capacity. On the Edwards Plateau of Texas, much of the increase in woody plants is due to an altered fire regime through the reduction in natural occurring wildfires since settlement (Fuhlendorf et al. 1996). Historically, these fires maintained a fairly open grassland by differentially enhancing establishment and growth of the non-woody plants and periodically killing fire-sensitive woody plants, such as Ashe juniper. These fires were dependent upon sufficient fine fuel and appropriate environmental conditions.

In addition to changes in the abundance of woody plants, there have also been major changes in the herbaceous component (grasses and forbs)(Fuhlendorf 1996). At the turn of the century stocking rates were severely high on most rangelands resulting in almost complete removal of vegetation cover. Some plant species are more resistant to grazing than others, because they can either tolerate repeated defoliation or are less preferred by livestock. As heavy grazing continued, the abundance of less resistant species decreased and frequently became locally extinct. Most of these species are the taller more productive species within the plant community, which results in the reduction of herbaceous productivity. As grazing caused these changes in species composition to less productive short-grasses, the likelihood of fire decreased because less fuel was available. Therefore, grazing interacts with fire to contribute to an increase of woody plants on historically fairly open rangelands by reducing the fuel required for fires. Changes in the herbaceous component of these grasslands appear to be reversible if stocking rates are reduced to moderate levels and pastures rested periodically (Fuhlendorf 1996).

In the past, grazing has also been thought to contribute to increases in woody plants by reducing the competitive ability of grasses to exclude woody plant establishment. However, recent research has suggested that this is not necessarily the case. Woody plants may require fairly specific environments to establish. Preliminary work (Smeins and Fuhlendorf, this volume) suggests that removal of herbaceous vegetation through grazing may result in a harsh environment where conditions may actually limit woody plant survival. Also, juniper seedlings produce less secondary chemicals than mature trees making them more palatable to livestock. So, grazing may not enhance woody plant establishment and may actually inhibit germination and establishment. This hypotheses is supported by several areas where grazing has been excluded for long periods of time (50-100 years) and woody plant densities have increased dramatically. The specific conditions required for germination and establishment of junipers are difficult to determine, but it appears that periodically the appropriate conditions result in substantial establishment under all grazing treatments over the past several decades.

Because changes in rangelands are caused by factors that are confounded and occur over a long time frame (100 years), they are difficult to evaluate through traditional field oriented experiments or observations by any one individual. Over the past 50 years, research at the Sonora Research Station, as well as other areas, has focused on the relationship between grazing, fire and vegetation change. We have utilized these studies to create a simulation model that allows the evaluation of long-term vegetation change and the options available to enhance long-term land management. Our objective is to utilize vegetation ecology to evaluate management scenarios for their long-term sustainability under variable weather patterns.

Model Description

This symposium focuses on two species of juniper. Because of differences between the species, the model was limited to the fire-sensitive, Ashe juniper. Analyses of vegetation change was approached as two relatively separate components at two different scales: 1) the landscape where the woody-herbaceous interaction occurs mediated by fire, and 2) the influence of grazing on herbaceous (non-woody) patches between woody plants (Fig.1). As woody plants (Ashe juniper) increase, the area dominated by herbaceous species decreases.

The woody plant component of the model represents the establishment of Ashe juniper from seeds and subsequent survival and growth through five size classes based upon canopy diameter (less than 0.75, 0.76-1.50, 1.51-3.00, 3.01-6.0 and greater than 6.0 m) (Fuhlendorf et al. 1996). State variables represent the density of trees (individuals/ha) with size class 3 (greater than 1.5 m or 6 ft canopy diameter) and larger trees representing seed producing trees and size class 5 (greater than 6.0 m or 24 ft canopy diameter) representing mature trees. Seedling establishment is a function of seed dispersal from off-site sources and eventually on-site seed production. Seed dispersal and establishment is represented as a stochastic event, similar to that expected by birds and mammals. On-site seed production is dependent upon the density of trees in the three largest size classes.

Tree mortality rates are dependent upon tree size/age and are functions of competition from other juniper trees and fire frequency and intensity. The influence of competition occurs when the densities of the two largest size classes reach critical levels. Mortality of trees through fire is dependent upon fire intensity which is influenced by the amount of herbaceous biomass and the season of occurrence. As trees increase in size and density they begin to influence the herbaceous biomass (Fuhlendorf and Smeins 1997) which decreases the effectiveness of fires. Three levels of fire intensity (high intensity = herbaceous biomass greater than 60%, medium intensity = herbaceous biomass between 30 and 60%, and low intensity = herbaceous biomass less than 30%) are evaluated for this symposium which are all cool-season fires. Maximum production of areas similar to the Sonora Research Station are approximately 3500 lbs/ac.

In addition to the woody component, a sub-model was developed of the influence of grazing on the herbaceous biomass. Grazing influences the herbaceous biomass directly by consuming forage and indirectly by altering the species composition. With heavy grazing, species composition of the herbaceous vegetation becomes dominated by short grasses, such as common curlymesquite and hairy tridens with herbaceous biomass varying between 350 and 1300 lbs/ac (10-40 %) depending upon weather, juniper size/density and periodic rest cycles from grazing. When grazing is removed, short-grasses decrease and there is an increase in more productive mid-grasses, such as Sideoats grama and Texas cupgrass resulting in production near the maximum (3500 lbs/ac). Moderate grazing, results in composition and production intermediate to the other treatments (1000 to 2600 lbs/ac. or 30 to 70%). Weather influences the rate of change. In addition to grazing and weather, biomass is decreased by the increase in size and density of juniper. So, these separate components are linked through herbaceous production and fire.

Parameters for these models were primarily determined from actual field observations and experiments (Smeins and Fuhlendorf, this volume). Several long-term studies of the influence of grazing were used to determine the changes in the herbaceous vegetation associated with different grazing intensities (Fuhlendorf 1996). Experimental studies were used to parameterize the influence of junipers on the herbaceous vegetation (Blomquist 1990, Fuhlendorf and Smeins 1997) and the influences of fire on different sized junipers (Wink and Wright 1975). The model was developed from over 50 years of data in an attempt to evaluate long-term scenarios. The starting conditions were considered an open grassland where any woody plants other than juniper are constant and not considered to be increasing. Also, the herbaceous composition was representative of a fairly heavily grazed landscape. The initial conditions are an attempt to simulate conditions resembling the historical changes that have occurred since settlement or the current situation of many Edwards Plateau rangelands following some kind of major control of existing Ashe juniper.

It is important that the accuracy of the model be considered under the following assumptions:

1. This model is parameterized for Ashe juniper on the western divide region of the Edwards Plateau. Its use on other junipers or on the moister canyonlands to the east would require different inputs.

2. Initial conditions represent an open grassland or at least a condition where woody species other than juniper are not increasing at a significant rate.

3. All fires and associated tree mortalities are based upon cool-season fires and only dependent upon the level of herbaceous fuel.

4. Fuel loads are considered continuous across the landscape and evaluation of tree density and establishment were not dependent upon spatial aspects of the vegetation.

5 The grazing intensity is based on forage utilization, so when herbaceous forage decreases the stocking rate must also decrease to prevent increasing the grazing pressure.

6. The influence of grazing was based primarily on cattle and sheep grazing, with minimal influence of goats. Goats consume a higher portion of woody plants so the results of this model would change dramatically under high levels of goat browsing.

7. The model uses a one-year time step and is intended to take a long term-perspective on the primary driving influences associated with the rangeland dynamics that contribute to the historical reduction in carrying capacity.

Simulated Influence of Grazing, Weather and Fire

Simulations were conducted on three grazing intensities (heavy, moderate and ungrazed) and three cool-season fire intervals (no fire, ten-years and twenty-years). For moderate and heavy grazing, simulations also included different scenarios where rest prior to fires was varied (no rest, one-year and two-year rest prior to burning). Data is presented as a series of graphs for each treatment over time with the following variables represented on the graphs: 1) potential herbaceous biomass (maximum production without grazing and under optimum weather patterns, could be considered the influence of the increase in juniper on herbaceous vegetation), 2) remaining herbaceous biomass (actual non-woody biomass including the influence of grazing, weather and the increase in juniper; the amount of fuel available for fires), and 3) density/ha of large juniper trees (greater than 6 m or 24 ft canopy diameter).

When the model was run with no fires the results were similar for all grazing conditions. Ashe juniper increased to form a dense canopy woodland after about 70 years, where the density of large trees (greater than 6 m diameter) approached 70 trees/ha (dash-dot line) (Fig. 2). This development took 70 years because it took time for juniper to establish and grow to a mature size class. Prior to the increase in large tree density there were dramatic increases in smaller trees (not shown on graphs) and the overall increase in juniper of all sizes is inverse to the decrease in potential herbaceous biomass (dotted line). Potential herbaceous biomass (primary forage for cattle and sheep and fuel for fire) was reduced by 60-70%. There is little difference between potential and remaining herbaceous biomass because without grazing maximizing herbaceous production is only limited by weather patterns. So even without grazing, 70 years without a fire resulted in a change from an open grassland to a dense canopy woodland and very little herbaceous production. Forage production and livestock carrying capacities were reduced by as much as 70% over 60 years when fires are not a part of the system and juniper is allowed to increase.

Without grazing, cool-season fires can maintain a fairly open grassland with a cool-season fire frequency of 20 years or less which sustains greater than 95% of the potential herbaceous production over 150 years (Fig. 3). At an interval of 25 years, the response of the model is highly variable and grasslands are maintained for over 100 years, but frequently woody plants escape fires to form a woodland. This suggests that without grazing, a threshold is approached at a fire frequency of approximately 25 years. If the frequency between fires is greater than 25 years, woody plants become dominant and change back to a grassland requires more severe, expensive control measures, such as mechanical or herbicide. This is based on initial conditions with no juniper on the landscape. With an increase in the initial density of trees, more frequent fires would be required.

The previous scenarios represented conditions without grazing, but most of the Edwards Plateau has experienced heavy grazing over the past 100-150 years and many rangelands continue to be stocked heavily today. With heavy continuous grazing, even frequent cool-season fires (less than 10 years) can not maintain an open grassland and result in a dramatic increase in Ashe juniper (dash-dot line) over about 60-100 years (Fig. 4). When rest periods of one- and two-years were included in the model (grazing removed prior to the fire), the rate of change decreased but juniper still increased to form a woodland resulting in a decrease in potential herbaceous production (dotted line) by 60-70% after 150 years (Fig. 5).

Historically, heavy grazing contributed to the formation of a closed-canopy woodland by reducing fuel loads and reducing the probability and intensity of fires. As the density and size of juniper increases, herbaceous production decreases (Fuhlendorf and Smeins 1997). This results in lower fire intensities which reduces mortality of junipers and eventually the carrying capacity of the rangeland is reduced. So, when rangelands are heavily grazed and fires occur at a 10 year return interval, livestock carrying capacity is reduced to about 40% of the original production (potential herbaceous biomass, dotted line), similar to conditions where fires were not present. Management of juniper through fire under heavy grazing is difficult and likely unsuccessful, although it can slow the rate of increase.

Under moderate grazing with no rest prior to fires, a 10-year fire frequency rarely maintains an open grassland (Fig. 6). The increase in woody plants is slightly delayed when compared to treatments that were unburned or had heavy grazing. After 80-100 years juniper has gained dominance (dash-dot line) and potential herbaceous production (dotted line) has decreased to about 40 % of the initial conditions. When a 1-year rest from moderate grazing occurs prior to each fire several alternative results can occur depending upon the weather variation (Fig. 7 and 8). The majority of the time (ca. 80%), density of large trees are maintained at less than 5 trees/ha and potential herbaceous production never drops below 90% of the original (Fig. 7). However, occasionally (ca. 20% of the time), random weather variation results in several low intensity fires allowing an increase in woody plants to form a closed canopy woodland after 100-120 years (Fig. 8). Therefore this treatment suggests that under moderate grazing long-term maintenance of grassland production may approach a threshold with a fire frequency near 10-years that is dependent upon the influence of weather variation on herbaceous production. A 1-year rest is necessary to produce enough fuel to consistently carry a fire. Otherwise, more frequent fires are required.

The 20-year fire frequency that maintained an open grassland in an ungrazed situation was not effective when any level of grazing occurred. Rest periods of 1- and 2-years prior to the fire, lengthened the time frame of the changes and suggests that if other control methods are incorporated, such as goats, a 20-year frequency could successfully maintain forage production.

Movement from a grass dominated plant community to one dominated by woody plants is associated with an interaction between fire frequency, grazing regime and weather variation. Figure 9 illustrates the average (n=20) response of the density of large (greater than 6 m diameter) trees/acre across grazing regimes, after 150 years with a fire every 10 years. Vertical bars represent the standard deviation. Large standard deviations suggest that variable weather can lead to several different states after 150 years. Under heavy grazing or moderate grazing with no rest prior to fires, a 10 year fire frequency will not maintain herbaceous dominated plant communities and result in a dramatic decrease in forage production and livestock carrying capacity.

Under most situations the initial starting conditions of an open grassland (with no junipers) are not realistic. In most cases, management practices, such as prescribed fires, are delayed until the problem is perceivable. This often means the density of trees has reached critical levels. To demonstrate the importance of applying management treatments prior to perceiving the problem, we applied the 10-year fire treatment to moderate grazing with a 1-year rest prior to the fire but delayed the first fire to occur on year 20, 30 and 40. So, these treatments skip the first, first two, and first three fires respectively. Figure 10 demonstrates that when the first fire is applied at year 10 under moderate grazing, densities of large trees are usually maintained below ten trees/ha after 150 years, depending upon weather variation prior to the fires. At the time of the first fire (year 10), all junipers are less than 0.75 m (about 2 ft.) canopy diameter (seedling and saplings) and would not appear to be an important component of the vegetation. When the first fire is delayed 10 years (year 20), the average density of large trees after 150 years is still relatively low but the variation is greater. If the first fire is not applied until year 30, the variation is greatly increased and more often than not the system becomes a closed canopy woodland after 70-100 years.

By delaying the initial fire to greater than 20 years, some trees are allowed to reach a critical size (1.5-3.0 m or over 6 ft. diameter) before the first fire. This causes a reduction in the herbaceous production (fuel load) reducing the effectiveness of fires and allows trees to reach the age of seed production further enhancing their rate of increase. Many areas of the Edwards Plateau, currently have junipers at a size and density greater than those associated with the conditions at year 30-40. This suggests that for many areas other techniques, such as mechanical, chemical or summer fires followed by periodic cool-season fires would be required to produce and maintain a grassland dominated community and livestock carrying capacities. Also, for landowners that are not currently experiencing an increase in juniper but the potential exists, pro-active management should begin prior to the onset of the issue.

Discussion

This model was based upon field data and illustrates the importance of a long-term ecological perspective in land management. Generally, Edwards Plateau rangelands are dominated by two states separated by a threshold. A grassland dominated state can be maintained by periodic fires that require grazing management to maintain herbaceous fuel loads. Without fires or if their severity is reduced because of low fuel, Ashe junipers establish and grow to a nearly closed-canopy woodland reducing the area dominated by herbaceous vegetation. This increases forage utilization on the remaining areas and lowers livestock carrying capacity of the rangeland. The threshold exists because once juniper density and size (about 5-6 ft) reaches a certain level, it is difficult for the vegetation to return to grassland dominance and reversing the trend may require more intensive management practices, such as mechanical methods (bulldozing, chaining, etc.) or more severe warm-season fires. These techniques may be dangerous, expensive and/or permanently destructive on areas with high potential for erosion.

Over-utilization by livestock is frequently considered a primary cause of the changes observed on Texas rangelands over the past 100 years. However, a dramatic reduction in carrying capacity is largely influenced by the increase in woody plants, such as Ashe juniper, which can be directly linked to the control of natural occurring fires. The primary direct influence of heavy grazing by cattle and sheep is the decreased abundance of taller, more productive mid-grasses, in favor of less productive, short-grasses which may also lower the carrying capacity. However, removal or reduction of stocking rates can often reverse these changes and return herbaceous production to previous levels within a grassland. An indirect influence of heavy grazing can be associated with the increase in Ashe juniper. Reductions in herbaceous productivity and heavy utilization of forage causes a reduction in fuel available for natural occurring or man-induced fires.

As juniper increases on rangelands, the carrying capacity is reduced to less than 40% of the maximum prior to the increase in juniper. So, essentially a rancher has reduced his potential earnings from livestock production by greater than 60%. Landowners that rely on livestock for their primary income are faced with a difficult challenge. They must maximize livestock production, while minimizing this decrease in carrying capacity associated with over-utilization and increased abundance of woody plants. The most economically and ecologically feasible management of fire-sensitive woody plants is periodic fires with a rotation of moderately stocked livestock that is synchronized to maximize herbaceous production prior to fires. The success of this approach requires initiation of fire treatments before a serious problem is observed and awareness of weather influences on herbaceous production. If an unsuccessful fire occurs, additional measures would be required to maintain grassland dominance and livestock production.

All simulated scenarios were based upon grazing by cattle and sheep with little or no influence from domestic browsers, such as goats. When goats are stocked at moderate to heavy rates, they consume many juniper seedlings, decrease growth rates of small trees and create browse-lines on large trees (Fuhlendorf 1992). All of these can slow the rate of increase in woody plants and decrease some of the negative influences the junipers have on herbaceous production. However, everything has a cost. Goats prefer many browse species over juniper and many of these species are important for wildlife, such as white-tailed deer. Therefore, continuous heavy utilization by goats can lead to reduction or elimination of important browse species as well as highly nutritious forbs and grasses. Goats have been extensively stocked on some Edwards Plateau rangelands and have frequently maintained fairly open grasslands. Removal of incentive payments for angora goats may lead to reductions of goat densities throughout Texas that will likely allow further increases in woody plants. Even without incentives, goats should be considered as potential technique for juniper management, but they should be used judiciously to minimize harmful effects on palatable secondary species that may be important to wildlife habitat.

In addition to altering the kind of animals, other approaches can facilitate management of juniper. Mortality of juniper from fires was dependent only on herbaceous fuel available at the time of fire and considered constant within that level of herbaceous fuel. In reality, high mortality to junipers can be achieved at lower fuel levels by conducting fires under more severe temperatures. Simulated fires were representative of cool-season burns conducted under conditions where fires were the safest to control. Fires conducted during warmer, drier weather can achieve higher mortality of trees, as well as account for lower levels of fuel or longer fire frequencies. These fires are more dangerous and require more skill to conduct. If restricted to the herbaceous layer fires can be utilized to control fairly high densities of large trees and still be relatively safe. However, if these fires become 'crown-fires' the potential for loosing control is greatly increased.

There are many scenarios for land management on these rangelands. If a closed-canopy woodland is desired, grazing intensity is irrelevant and could possibly enhance the control of fires that would result in an increase in woody plants, such as Ashe juniper. However, if the objective is livestock production or other enterprises that require maintained herbaceous production, then it is essential to manage juniper. Several potential management strategies will maintain grassland production depending upon the objectives and expertise of the landowner. Cool-season fires will maintain grassland production but will not contribute to restoration from woody domination. The success of fires depends on the grazing intensity, rest from grazing prior to fires, fire frequency, and weather patterns. Goat browsing, warm-season fires and intensive management practices, such as bulldozing, chaining, and herbicide treatment can be used to enhance management or restore grass production to areas that have already crossed critical thresholds.

Conclusions

1. Reduction in natural occurring fires is the primary cause of the increase in fire-sensitive woody plants, such as Ashe juniper. The success of a fire is dependent upon the availability of fuel and environmental conditions at the time of fire. Cool-season fires require high levels of herbaceous biomass, while warm-season fires can occur in the crown of dense canopy woodlands which can be dangerous but effective.

2. Regardless of grazing intensity, when fire is limited many areas on the Edwards Plateau experience an increase in Ashe juniper that eventually causes a decrease in livestock carrying capacity. This decrease begins about 20-30 years after the first juniper plants appear and continues over the next 30-40 years until the livestock carrying capacity has decreased by 60-70%.

3. Increase in juniper density and size is exponential. Initial establishment and growth is slow, but at a critical threshold (ca. 25-50 years) the rates increase until a dense-canopy woodland is formed (60-80 years). Many rangelands may already be beyond this threshold and require higher input management techniques.

4. Grazing is the primary factor that influences species composition of herbaceous vegetation. This influences herbaceous production because some species are more productive than others. Grazing, also removes a major portion of existing forage, regardless of the herbaceous composition. Herbaceous production is the link between grazing and the maintenance of grasslands through fire.

5. The maintenance of a relatively open grassland for 100-150 year, requires cool-season fires every 20 years for an area that is not grazed and every 10 years for a moderately grazed area. Under any level of grazing, a rest prior to fires increases the success of juniper management by increasing herbaceous fuel.

6. Under heavy grazing, fires are not successful because of low fuel levels. If sufficient rest from grazing occurs prior to fires, they can delay the process and maintain a fairly open grassland for about 60-100 years.

7. In addition to fire frequency, the timing of the first fire is important because after about 30 years trees reach a critical abundance and significant size to reduce herbaceous production and begin an exponential increase. Maintenance of grasslands requires pro-active management (initiation of fires prior to observation of juniper).

8. Weather patterns interact with grazing intensity and fire to influence the vegetation. Periodic above average rainfall and droughts influence the herbaceous production which essentially alters the grazing intensity under constant stocking rates. Variable levels of production influence the amount of vegetation capable of carrying a fire across the landscape.

9. Goats can be used to lengthen time intervals between fires and reduce the negative influences of juniper on herbaceous vegetation. Warm-season fires can be used to increase mortality to larger trees, lengthen time intervals between fires and counteract low levels of herbaceous fuel. Each of these alternatives is associated with additional costs.

10. Sustainable livestock production on many rangelands in Texas requires management of juniper. This can be achieved through proper grazing management and periodic fires.

References

Blomquist, K.W. 1990. Selected life history and synecological characteristics of Ashe juniper on the Edwards Plateau of Texas. M.S. Thesis, Dep. Rangeland Ecology and Management, Texas A&M Univ., College Station, TX.

Fuhlendorf, S.D. 1992. Influence of age/size and grazing history on understory relationships of Ashe juniper. M.S. Thesis, Dep. Rangeland Ecology and Management, Texas A&M Univ., College Station, TX.

Fuhlendorf, S.D. and F.E. Smeins 1997. Browsing and tree size influence on understory of Ashe juniper. Journal of Range Management (in press).

Fuhlendorf, S.D. 1996. Multi-scale vegetation responses to long-term herbivory and weather variation on the Edwards Plateau, Texas. Ph.D. Dissertation, Dept. Rangeland Ecology and Management, Texas A&M University, College Station.

Fuhlendorf, S.D., Smeins, F.E., and W.E. Grant. 1996. Simulation of a fire-sensitive ecological threshold: a case study of Ashe juniper on the Edwards Plateau of Texas, USA. Ecological Modelling 90:245-255.

Wink, R.L. and H.A. Wright 1975. Effects of fire on a juniper community. Journal of Range Management 26:326-329.