«Summer Establishment of Sonoran Desert Species for Revegetation of Abandoned Farmland Using Line Source Sprinkler Irrigation BRUCE A. ROUNDY ...»
Block and block 3 irrigation e†ects on plant density were generally not signiÐcant, except for some sample dates (Table 2). SigniÐcant block e†ects on grass density and block 3 irrigation e†ects on shrub density for the earlier sampling dates in 1992 may have been associated with uneven water applications. Although irrigation was conducted at night to minimize wind drift, some drift was unavoidable. Uneven application of water was also associated with imperfect leveling of the Ðeld, but there was no surface crusting and little runo† or run-on was observed. Applied water varied most at the greatest distance from the line source (Figure 1). Water applications were more variable in 1992 than 1993. Coefficients of variation for individual days of irrigation at di†erent distances from the line source in 1992 averaged 9.4% at 1.5 m, 8.8% at 4.5 m, 11.4% at 7.5 m, 27.3% at 10.5 m, and 108.6% at
13.5 m. In 1993 they averaged 3.2% at 1.5 m, 1.8% at 4.5 m, 1.3% at 7.5 m, 2.2% at
10.5 m, and 3.4% at 13.5 m. Variation can be reduced by irrigating only during windless times. Overall, the line source sprinkler produced a gradient in water B. A. Roundy et al.
28 FIGURE 1 Irrigation and precipitation at six distances from a line source sprinkler in 1992 and 1993. (Error bars indicate 95% conÐdence intervals.) applied that resulted in di†erences in plant establishment, but the greatest di†erence in water availability and plant establishment was between irrigated and unirrigated soils.
Species, irrigation, and the species 3 irrigation e†ects on plant density were signiÐcant (p, 0.05) for most sample dates for grasses and woody species (Table 2).
Bursage, Ambrosia deltoidea, had limited emergence and was not included in the analysis. ProÐle contrasts indicated that, in general, density of both grasses and woody plants varied mainly among the lower levels of irrigation and nonirrigated areas, rather than among the higher levels of irrigation. Species interactions with plant density di†erences associated with di†erent irrigation levels were also mostly signiÐcant (p, 0.05) under less or no irrigation, than under higher irrigation.
Sonoran Desert Revegetation 29 FIGURE 2 Average daily air temperature, relative humidity, and date and amount of summer rainfall and irrigation from a line source sprinkler during the summers of 1992 and 1993 at Tucson, Arizona. (Stacked bars represent the amount of irrigation water applied at the indicated distance from a line source sprinkler. Open bars indicate amount of rainfall.) Grasses In 1992 and 1993, most grass species sown within 10.5 m of the line source reached maximum density two weeks after the Ðrst irrigation and stayed near that density into September (Figures 4 and 5). In general, most grasses established well at a B. A. Roundy et al.
30 FIGURE 3 Soil water potential at four distances from a line source sprinkler for Ðve soil depths during the summers of 1992 and 1993 at Tucson, Arizona.
FIGURE 4 Plants per m of row of six grasses sampled at seven distances from a line source sprinkler on 12 dates in summer 1992 at Tucson, Arizona.
require two weeks of available water to initiate adventitious roots which arise near the soil surface (Roundy et al. 1993). These grasses are susceptible to postgermination dry periods and may have high mortality if rainfall or irrigation sufficient to germinate seeds is followed by 1È2 weeks of drying (Roundy, Abbott, and Livingston 1997). In this study, germination and seedling emergence of grasses were much more sensitive to the irrigation gradient and di†erences between irrigation and rainfall than was seedling survival. The consistency of initial irrigation and natural rainfall, as well as the initially wet subsurface soil in this study, combined to allow survival of seedlings once they emerged.
The grass species di†ered in their sensitivity to irrigation and rainfall. Although Arizona cottontopÈDigitaria californica, had the highest density of any of the Sonoran Desert Revegetation 33 FIGURE 5 Plants per m of row for six grasses sampled at seven distances from a line source sprinkler on Ðve dates in summer 1993 at Tucson, Arizona.
grasses under irrigation, it had limited establishment with no irrigation (Figures 4 and 5). Spike dropseed had much higher emergence in 1992 than 1993. Lehmann lovegrass exhibited a late germination pattern, with limited emergence from irrigation but acceptable establishment from summer rains under no irrigation in 1992 (Figure 4). This species also emerged later than the other grasses in 1993, but established better with irrigation than with rainfall alone in that year (Figure 5).
Although Lehmann lovegrass has a reputation for much more dependable establishment than native grasses in rangeland seedings (Roundy and Biedenbender 1995 ;
Roundy et al. 1996), it did not establish as well as the native grasses in this study.
All of the native grasses showed good potential to establish under irrigation for these soils, soil moisture conditions, and climate. All of the grasses except galleta are B. A. Roundy et al.
34 known to have the C4 photosynthetic pathway (Waller and Lewis 1979). Cane beardgrass required the least water for high and moderate establishment, and had low establishment only with the least amount of precipitation (Table 3). Purple threeawn and spike dropseed had moderate establishment with only 76 mm of precipitation. While most of the grasses had moderate establishment with 76È210 mm of irrigation or rainfall, consistent establishment would probably require at least 200 mm (Table 3). Success would be maximized by sowing prior to summer rains in July, then subsequently irrigating daily for 1È2 weeks to keep the soil wet until seedlings emerge. The 200 mm would wet the sandy soil in this study to approximately 30È40 cm deep which should be sufficient to encourage sustained root growth. If initial irrigation and emergence were not followed by summer rains, additional irrigation would help adventitious roots develop and ensure Ðrst-season establishment. Successful establishment of these grasses on sites with Ðner-textured soils would probably require more irrigation and rainfall due to lower inÐltration rates and slower emergence and root growth.
Shrubs and Trees ProÐle analyses indicated that di†erences in woody plant densities generally occurred at distances between 7.5 and 16.5 m from the line source. All species emerged within a week of initial irrigation except for jojoba which emerged about two weeks after the other species (Figures 6 and 7). In 1992, all species except creosotebush had good emergence (. 10 plants per m of row) at a distance of 10.5 m or closer from the line source (Figure 6). Irrigation plus precipitation at these distances was greater than or equal to 210 mm (Figure 1). Most species had lower seedling densities in 1993 than 1992, and in that year seedling densities were also highest within 10.5 m of the line source, equaling 187 mm or more of irrigation plus precipitation (Figures 1, 6, and 7). Seedling survival of woody species was generally higher
FIGURE 6 Plants per m of row for six woody species sampled at seven distances from a line source sprinkler on 12 dates in summer 1992 at Tucson, Arizona.
in 1992 (. 90%) than in 1993 (about 50È60%). There were 14 major irrigations in 1992 and 11 in 1993 (Figure 2), and even the surface of irrigated soils dried out sooner in 1993 than 1992 (Figure 3). This may have a†ected the seedling survival of woody species in 1993, but grasses during the same time period had limited mortality.
In general, mesquite and palo verde required the least irrigation and precipitation (about 107 mm) for consistently successful establishment (. 5 plants per m of row, Table 3). Cat claw acacia had greater establishment in 1992 than 1993 (Figures 6 and 7). Creosotebush required more irrigation plus precipitation to establish successfully than did any of the other woody species (Table 3). All species but creosotebush had some establishment with no irrigation on both years, although B. A. Roundy et al.
36 FIGURE 7 Plants per m of row for six woody species sampled at seven distances from a line source sprinkler on Ðve dates in summer 1993 at Tucson, Arizona.
jojobaÏs establishment was only 0.6 plants/m of row under natural rainfall in 1993 (Figures 6 and 7). As with the grasses tested, 200 mm of irrigation plus precipitation should be sufficient to establish most of these species under similar conditions (Table 3). To bring the upper 60 cm of this soil to Ðeld capacity water content would require 300È500 mm of irrigation plus precipitation, depending on evaporation.
Tap root depths of acacia, mesquite, and palo verde were greater than 50 cm deep by October of 1992 and November of 1993, while those of jojoba were less than 35 cm deep. Tap root depths for mesquite and acacia increased with distance from the line source, while those of palo verde and jojoba decreased. For example, tap root depths in November 1993 were 65 and 75 cm deep for acacia at 1.5 m and
13.5 m from the line source, those for mesquite were 85 and 113 cm, those for jojoba Sonoran Desert Revegetation 37 were 48 and 31 cm, and those for palo verde were 84 and 65 cm. Irrigation of 200È300 mm, in addition to natural precipitation should allow these species to germinate, while Ðlling up the soil proÐle sufficiently to permit sustained root growth and survival.
Although some of these woody species established with 107 mm of irrigation plus precipitation, shoot growth was much greater for acacia, mesquite, and palo verde when total water received was over 265 mm. Shoot heights in October 1992 for plants receiving at least 265 mm total water were 43.2 cm for acacia, 46.7 cm for mesquite, and 54 cm for palo verde. For plants receiving only 107 mm total water, heights were 15.5 cm for acacia, 18.3 cm for mesquite, and 34.7 cm for palo verde.
Jojoba had much less shoot and tap root growth than the other woody species measured and is known for its slow growth (Roundy and Ruyle 1989), and also for susceptibility to drought at the seedling stage (Sherbrooke 1976). Jojoba seedlings in this study elongated tap roots faster than shoots. In early September (11 September 1992, 72 days after 1 July) tap roots were 20 cm deep, while main shoots were only about 5 cm tall. Jojoba shoot growth was not increased by increased irrigation.
Shoot heights in October 1992 associated with total water received were 6.6 cm for
356.7 mm water, 7.1 cm for 265 mm water, and 7.5 cm for 107 mm water. Water applications to produce high shoot growth could also produce excessive densities for some woody plants in a semiarid environment. If such high levels of irrigation are used (. 265 mm), pure live seeding rates should be reduced.
Discussion The line source sprinkler was useful in determining water requirements for establishment of the herbaceous and woody species tested. Although a gradient in irrigation was created by the line source, it did not result in a signiÐcant measurable di†erence in soil water potential and availability within the irrigation gradient. In this study, the soil proÐle was initially at Ðeld capacity prior to sowing and before initiation of the line source irrigation. Using the line source on initially dry soils would create a greater gradient in soil water potential and availability.
Periods of irrigation, natural rainfall, and soil water availability at the 1È3 cm depth on the irrigated and unirrigated soils were associated with seedling emergence patterns. Most species emerged on irrigated soils within a week of irrigation and did not emerge on unirrigated soils until the onset of summer rainfall. Exceptions to the rapid emergence pattern were Lehmann lovegrass and jojoba, both of which had delayed emergence. Lehmann lovegrass emerged best after summer rainfall and established only on nonirrigated soils in 1992. This may indicate that irrigation was sufficient to germinate, but not to maintain high soil water availability at the extremely shallow depths of emergence for this species (Winkel, Roundy, and Blough 1991). On the other hand, jojoba was slow to emerge, but eventually established best when irrigated, indicating that it started to germinate with irrigation but had slow shoot growth. Lack of di†erences in surface soil water availability measurements under the irrigation gradient did not explain di†erences under the gradient in grass density. Availability of water to seeds may have varied with the irrigation gradient on a smaller spatial scale than the 1È3 cm depth measured by the moisture sensors. It has been difficult to measure the moisture microenvironment at the scale of small seeds (Call and Roundy 1991).
Although data from this experiment suggest that some species will establish with less water than others, the most important Ðnding is that most of the species tested will establish with direct seeding and irrigation. A possible strategy for revegetating abandoned farmlands with similar soils and climatic conditions would be to irrigate daily for 1È2 weeks after sowing in July until seedlings emerge, applying a total of 200È300 mm of water. This should be sufficient to Ðll the soil proÐle B. A. Roundy et al.
38 enough to sustain seedling survival, especially with additional water from summer rainfall. Direct seeding of these species without supplemental irrigation is risky due to the erratic and limited nature of summer rainfall. From a practical standpoint, revegetation of these lands will be most successful if done right after abandonment, before the irrigation system falls into disrepair. For lands without an irrigation system, these plants show high potential for establishment if microcatchments (Edwards et al. 2000) or spot irrigation devices (Bainbridge and Virginia, 1990) can supply sufficient water for germination and root growth.
References Allen, E. B., and L. L. Jackson. 1992. The arid west : here the challenge is to reverse the downward spiral of desertiÐcation over vast areas. Restoration and Management Notes 10 : 56È59.
Arizona Meteorological Network. 2000. http ://ag.arizona.edu/azmet Bainbridge, D. A., and R. A. Virginia. 1990. Restoration of the Sonoran Desert of California.
Restoration and Management Notes 8 : 3È14.
Call, C. A., and B. A. Roundy. 1991. Perspectives and processes in revegetation of arid and semiarid rangelands. Journal of Range Management 44 : 543È549.