Multidecadal Drought Cycles in South Central New Mexico Patterns and Consequences

Water Freedom System

Survive Global Water Shortages

Get Instant Access

Bruce T. Milne Douglas I. Moore Julio L. Betancourt James A. Parks Thomas W. Swetnam Robert R. Parmenter William T. Pockman

Extreme, regional droughts are the most common form of disturbance in semiarid ecosystems typified by relatively slow recovery rates. Drought-driven impacts can include regionally synchronized insect outbreaks, wildfires, and tree mortality (Swetnam and Betancourt 1990), as well as disastrous failures of agriculture, silviculture, and livestock production (Mainguet 1994). Drought conditions, accompanied by anthropogenic land mismanagement, have led to subsequent invasions of grasslands and farmlands by woody shrubs and nonna-tive forbs and grasses, contributing to the modern "desertification" process manifested in many parts of the world (Archer et al. 1988).

In the American Southwest, the drought of the 1950s was one of the most severe climate events of the past millennium because of wide ramifications for the region's ecology (Herbel et al. 1972; Swetnam and Betancourt 1998), water resources (Thomas 1963), and economy (Regensberg 1996). As human population and resource needs increase in the Southwest, so will the economic sensitivity to large-scale drought. A clear understanding of extreme droughts is necessary not only to understand long-term ecosystem dynamics, but also to mitigate socioeconomic impacts.

The goals of this chapter are to use the Sevilleta LTER site in central New Mexico to (1) quantify the decadal variability in precipitation inferred from a 394-year record of tree rings, (2) relate the repeated decadal fluctuations in precipitation to major droughts of the 1890s and 1950s, (3) assess the ecological responses associated with droughts of the last century, and (4) elucidate the biotic-atmospheric feed backs that may influence future responses. We assess the magnitude, timing, and consequences of decadal fluctuations in annual precipitation.

Sevilleta LTER Site Description

The Sevilleta LTER research site is located at the Sevilleta National Wildlife Refuge (NWR), Socorro County, New Mexico (34°20' N, 106°50' W). The Sevilleta NWR comprises 100,000 ha of grassland, desert, and woodland bordered by two mountain ranges and the Rio Grande Valley in between. Elevations range from 1,350 m at the Rio Grande to 2,797 m at Ladrón Peak in the northwestern portion of the refuge. Topography, geology, soils, and hydrology, interacting with major air mass dynamics, provide a spatial and temporal template that makes the region a transition zone between several biomes. The region contains communities that both represent and intersect Great Plains Grassland, Great Basin Shrub-steppe, Chi-huahuan Desert, Interior Chaparral, and Montane Coniferous Forest (Brown 1982).

Sevilleta Climate Description

The Sevilleta LTER study region straddles the boundary between major seasonal air masses (e.g., winter arctic frontal systems descend southward across the Great Plains and influence Sevilleta's eastern edge; Great Basin polar air masses extend to Sevilleta's northern edge; the Bermuda High generates summer convective storms over the mountains, which track northeast across Sevilleta's lowlands). Superimposed on these spatial patterns are the temporal dynamics of the El Niño-Southern Oscillation (ENSO) phenomenon. These climate phenomena are further translated by the orography of the southern Rocky Mountains.

The annual climate of the Sevilleta area includes two different patterns of seasonal storms. During the late fall, winter, and spring, storms develop in both the northeastern and tropical Pacific Ocean. These storms are steered into the region by both the polar and subtropical jet streams, and their annual frequencies and total precipitation over central New Mexico are in part modulated by both interannual (ENSO) and decadal-scale (Pacific Decadal Oscillation) variability in Pacific climate. In general, central New Mexico tends to have wet falls, winters, and springs during El Niño events, or times when the Pacific Decadal Oscillation (PDO) is positive, and the opposite conditions during La Niña (for ENSO, see Andrade and Sellers 1988, Molles and Dahm 1990, Redmond and Koch 1991, Cayan and Webb 1992, Kahya and Dracup 1993; for PDO, see Cayan et al. 1998, Mantua and Hare 2002). At Socorro, just south of the Sevilleta, precipitation during the period from October through May increased by 53% during El Niño years and decreased by slightly more than half during La Niña events when compared to medial years over the past 80 years (figure 15.1, table 15.1). Although this fall-winter-spring precipitation increase may appear small, the colder conditions and dormant vegetation favor soil moisture recharge, thereby increasing soil water availability for the growing season. Also, most of these cool-season frontal storms have broad tracks

Precipitation mm 600:

I ' 1 1 1 I ' 1 1 ' I ' 1 ' 1 I 1 ' 1 1 I 1 1 1 1 I 1 1 1 1 I 1 1 1 1 I 1 1 ' 1 I 1 1 ' 1 I ' 1 ' 1 I 1 1 1 1 I 1 1 1 1 I

1860 1890 1900 1910 1920 1933 1940 1950 1930 1970 1980 1990 2000


I ' 1 1 1 I ' 1 1 ' I ' 1 ' 1 I 1 ' 1 1 I 1 1 1 1 I 1 1 1 1 I 1 1 1 1 I 1 1 ' 1 I 1 1 ' 1 I ' 1 ' 1 I 1 1 1 1 I 1 1 1 1 I

1860 1890 1900 1910 1920 1933 1940 1950 1930 1970 1980 1990 2000


Figure 15.1 Precipitation record for Socorro, New Mexico (1880-2000). Note periods of low precipitation from the late 1890s to early 1900s and during the period 1950-1957.

(>100 km) and tend to synchronize ecosystem response to precipitation events across the region.

Summertime precipitation is driven by the North American Monsoon. During the summer, the jet stream retreats to more northern latitudes, thereby reducing the influence of Pacific storm systems on the region. Normally, the period of greatest precipitation in much of the Southwest occurs during the months of July, August, and September, and is associated with convective thunderstorms during this North American Monsoon. Monsoon precipitation originates from moist air masses over the Gulf of Mexico and is pushed into the Southwest by the Bermuda High (Mitchell 1976; Neilson 1986). The resulting precipitation is distributed heteroge-neously on the landscape by thunderstorms that originate over the mountains and move over the lowlands. Although temporal variability is low, spatial variability is

Table 15.1 Mean precipitation amounts by ENSO class from Socorro, New Mexico, 1914-1993a.

Precipitation (mm)


class N Annual Oct-May Jun-Sep

La Niña 9 162.5 b 49.9 c 112.5 a a ENSO classes with different letters are significantly different (P < 0.05).

high. The linkage between the ENSO phenomenon and summer precipitation in New Mexico is weak (Andrade and Sellers 1988; Molles et al. 1992; Harrington et al. 1992). There seems to be some predictability of the onset and/or magnitude of the monsoon moisture (Higgins and Shi 2000), but the predictability is much poorer than for the winter and spring moisture associated with extremes of the ENSO cycle. Although the monsoon moisture of July, August, and September generally accounts for well over half of the annual precipitation in the Sevilleta region, high evapotranspiration prevents summer moisture from recharging soil moisture to depth. Thus, vegetation responses to summer moisture are ephemeral and patchy.

Therefore, we regard the fall-winter-spring precipitation as separate from that of the summer (June-September) precipitation. Continuous monitoring of soil moisture in the Sevilleta shows that the increased temperatures and plant growth of May and June deplete surface soil moisture from the nonmonsoon period before the monsoons begin in July. Monsoon moisture is generally lost quickly (2-3 days) between storms through evaporation, so that the transition from summer to winter is often a period of dry soils as well.

Multidecadal Precipitation Patterns

Direct measurements of precipitation at the Sevilleta LTER began in 1988, and records for Socorro County date only to the 1870s. Consequently, we elucidated persistent interdecadal weather patterns from a record of annual precipitation obtained from tree rings. A 394-year precipitation history (1598-1991) was reconstructed by Parks et al. (in press) by combining pinon pine (Pinus edulis) tree-ring chronologies from live and dead trees (killed by the 1950s drought) in Arroyo de Milagro in the Los Pinos Mountains, in the eastern sector of the Sevilleta NWR (figures 15.2, 15.3a). The original reconstructed precipitation depths are presented as deviations from the long-term mean (286 mm, s.d., = 63.5) to emphasize anomalous wet and dry years (figure 15.3a). Details of this particular reconstruction are explained in Betancourt et al. (1993) and Parks et al. (in press). The Sevilleta tree-ring index chronology was statistically compared to actual precipitation measured in Socorro for the period of overlap (1892-1991). We compared several seasons, but the highest correlation (r = 0.74) was for a 13-month interval covering the prior August through the current August of the growth year. For southern New Mexico conifers, Grissino-Mayer et al. (1997) found a similar conditioning of cambial growth by rainfall during the previous summer and fall and the current spring and summer, with winter rainfall (January through March) having little effect. Linear regression was used to estimate August-to-August precipitation from tree-ring indices.

We used multitaper method spectral analysis (MTM; Dettinger et al. 1995) to measure the temporal variability in the reconstructed precipitation record. MTM spectral analysis uses three orthogonal waveforms to measure the variance in a time series at a variety of frequencies, or inversely, periods. Confidence intervals (95 and 99%) were defined under the geophysically appropriate null hypothesis

Figure 15.2 Pinon-juniper woodlands in the Los Pinos Mountains, Sevilleta National Wildlife Refuge, New Mexico. Pinon pines in this mountain range provided samples for reconstruction of the long-term precipitation record.

that the time series is red noise, that is, more correlated than white noise and less correlated than a random walk. Thus, peaks in the MTM spectrum that exceed the confidence limits indicated persistent oscillations (figure 15.3b). At the decadal scale, the Sevilleta tree-ring reconstruction evidenced significant (99% confidence interval) variability for periods of 72.9, 68.5, 64.1, 60.2, 56.8, 53.8, 51.3, 48.8, 44.4, 42.7, 40.9, 39.4, and 36.6 years (mean = 51.9, s.d. = 11.3). Thus, on average, droughts and wet periods have occurred about every 52 years in Arroyo de Milagro. The data must be evaluated with caution because detrending during the construction of the original tree-ring record may have led to an overestimate of variation in the low frequencies (D. Meko, pers. comm., 2002).

Droughts present plants with conflicting demographic challenges. A seed is in a race between (1) the appearance of suitable conditions for germination and (2) the ever-diminishing chance of surviving or escaping predation. Droughts retard the former and may accelerate the latter. Thus, an ecologically relevant analysis of annual precipitation records should focus on the deviations from the mean precipitation because the deviations measure the uniqueness of any given year. Moreover, a string of wet years favors germination and establishment of species that require anomalously high amounts of water, such as the blue grama grass, Bouteloua gracilis (Neilson 1986). A string of dry years should exhaust seed banks through attrition and thereby color plant community dynamics for years to come. Thus, we recast the annual estimates of precipitation as deviations from the mean, or "normal" (figure 15.3a), and examined the strings of successive deviations to find epochs of persistent wet and dry conditions.

Figure 15.3 Patterns of long-term variation of precipitation based on the Arroyo de Milagro tree-ring record. (a) Annual precipitation depths minus the long-term mean. Dots indicate years that occurred in dry epochs. (b) Multitaper method spectrum of the raw precipitation depths. Horizontal lines indicate 95% (dashed) and 99% (solid) confidence intervals. (c) Running sum of the precipitation anomalies, that is, cumulative precipitation anomalies, CPA. (d) Cumulative probability distribution of the CPAs.

Figure 15.3 Patterns of long-term variation of precipitation based on the Arroyo de Milagro tree-ring record. (a) Annual precipitation depths minus the long-term mean. Dots indicate years that occurred in dry epochs. (b) Multitaper method spectrum of the raw precipitation depths. Horizontal lines indicate 95% (dashed) and 99% (solid) confidence intervals. (c) Running sum of the precipitation anomalies, that is, cumulative precipitation anomalies, CPA. (d) Cumulative probability distribution of the CPAs.

Following Feller (1968) and the technique of "mass curves" in hydrology (e.g., Peng and Buras 2000), we revealed persistent, repeated wet and dry periods by forming the running sum of the anomalies, that is, cumulated precipitation anomalies or CPAs (figure 15.3c). Peaks and troughs correspond to high and low net accumulations, respectively. The maximum CPA (787 mm) occurred in 1727, and the low of -523 mm occurred in 1971. Plants living in 1727 or 1971 would have been at the ends of extremely wet or dry epochs, respectively. The cumulative probability density of the CPA indicated that only 10% of the years were in major droughts like the early 1600s and 1950s (figure 15.3d). In contrast, the skewed distribution of anomalies produced wet epochs that exceeded one standard deviation of the CPA 20% of the time. Individual years were somewhat more likely to be in the midst of a wet epoch than a dry one.

There were 7 lengthy strings of dry years that constituted persistent droughts. Dividing the record length by 7 suggested 56 years as an approximate drought recurrence interval, which is in agreement with the time series analysis (figure 15.3b). The interval coincided with recent major droughts of the 1890s and 1950s that were about 60 years apart.

We concluded that long-term fluctuations in precipitation have occurred over the

Sevilleta region about every 52 years, although intervals of 41 to 63 years are within one standard deviation of the expected 52 years. Attempts to predict future droughts are limited by the considerable variability in the recurrence interval and the lack of a definitive theory about the climatic origin of the oscillation. However, this limitation does not deny the existence of decade-long droughts in the past that could recur with major ecological and socioeconomic consequences.

Possible Sources of Decadal to Multidecadal Climate Oscillations

The ultimate source of the low-frequency variations in the Sevilleta tree-ring series is presently unresolved. Oscillatory modes at decadal to century timescales have been identified in annual temperature and precipitation series from instrumental records (e.g., Cayan et al. 1998; McCabe and Dettinger 1999; Dettinger et al. 2001) and tree rings across western North America (e.g., Biondi et al. 2001; Gedalof and Smith 2001; Villalba et al. 2001). It is assumed that most of this low-frequency variation originates in the Pacific Basin and involves interaction of the ENSO mode with longer term decadal to centennial fluctuations in climate (Dettinger et al. 2001). Note that similar low-frequency variations have also been identified for the Atlantic Ocean (e.g., Delworth et al. 1993; Enfield and Mestas-Nunez 1999) and could produce similar periodicities in temperature and precipitation, particularly in summer. In the North Pacific, much of the sea surface temperature variance occurs in a mode with decadal (~20-30 years) timescales and is accompanied by variability in the strength and position of the Aleutian Low in winter. This has been called the North Pacific Oscillation (NPO; Gershnov and Barnett 1998) or Pacific Decadal Oscillation (PDO; Mantua and Hare 2002).

Among the leading explanations for low-frequency variations such as the PDO are stochastic atmospheric forcing, atmospheric teleconnections, midlatitude ocean-atmosphere interactions, tropical-extratropical interactions, oceanic telecon-nections, and intrinsic ocean variability (see summaries in Hare et al. 2000; Dettinger et al. 2001; Mantua and Hare 2002). There is considerable discussion about the steady state versus chaotic behavior of decadal- to century-scale variability, and thus about its predictability. An optimistic view is that knowledge about the present phase of the long-term mode (e.g., PDO) can be used to forecast climate several years ahead (Latif and Barnett 1996; Dettinger et al. 2000; Schneider and Miller 2001).

The twentieth century was marked by two full PDO cycles. The "cool" or negative PDO (more La Nina-like) regime prevailed from 1890 to 1924 and 1947 to 1976, whereas the "warm" or positive PDO regime prevailed from 1925 to 1946 and from 1977 to 1994. Using wavelet analysis, Minobe (1997, 1999) found that fluctuations in North Pacific sea surface temperatures, sea level pressures, and temperature reconstructions based on North American tree rings were most energetic at periodicities of 15-25 years (for boreal winter) and 50-70 years (for boreal winter and spring). According to Minobe (1999), the two periodicities synchronize with a relative period of three and produce a "regime shift" in North Pacific climate in the 1920s, 1940s, and 1970s when they reverse phase. There is a slight indication for such a regime shift starting in 1998 when the PDO index turned sharply negative ("cool" mode), the tropical Pacific began to cool after the prolonged post-1976 warming, and the North Atlantic started to warm after prolonged cooling since the 1960s. Such were the conditions during the regime shift of the 1940s that led into the 1950s drought, an episode that, according to Sevilleta tree rings, tends to recur every 41-63 years. We resist the temptation to make any long-term forecast, but we suggest instead that New Mexico politicians, resource managers, and ranchers alike have little cause for optimism. And as Swetnam and Betancourt (1998) have pointed out, given the possibility for another extreme drought, local biologists should be poised to take advantage of such "natural experiments."

Reliability of Arroyo de Milagro Reconstruction for Regional Assessment

Our analyses of the tree-ring precipitation chronology indicated that major droughts have occurred at Arroyo de Milagro in the Sevilleta every 41-63 years (figure 15.3b,c). Overlap between the tree-ring record and modern meteorological measurements enabled us to evaluate the extent to which the Sevilleta record indicates conditions over the broader Middle Rio Grande Basin.

Measured total annual precipitation can be a poor predictor of ecosystem productivity because the seasonal timing of precipitation is critical to its biotic effectiveness. Timing and other factors such as soil moisture holding capacity can influence the effect of the precipitation on growth (Valentine and Norris 1964). The Palmer Drought Severity Index (PDSI; Palmer 1965) can identify ecologically effective wet and dry conditions. The PDSI uses precipitation, temperature, and soil moisture to give a measure of moisture availability. The index is standardized to the local climate, so that relative dryness and wetness can be compared across the entire United States. The PDSI is extrapolated over climate regions within states.

The PDSI record for the Sevilleta region covers the entire period from 1895 to the present (figure 15.4; source drought/main.html). Periods when the index is above 0 are considered wet periods, whereas those below zero denote dry conditions. Drought classifications begin at -2.0 and increase in severity with decreasing index values. For the relatively short period of the Sevilleta LTER (1989-present), conditions have oscillated from the dry La Niña of 1989 to quite wet during the 1992-1993 El Niño. The index then stayed mostly dry until the wet monsoon of 1996, which carried on through the 1997-1998 El Niño and then back to mostly dry for the extended La Niña of 1998 to the present (2002). The wet to dry oscillation is obviously the norm.

The most noticeable exceptions to this pattern are during two extended periods during the 1890s to early 1900s and from 1950 to 1957. Then, the index stayed below zero for the entire time. The 1890s to early 1900s drought was responsible

Palmer Drought Severity Index

Palmer Drought Severity Index

Figure 15.4 Palmer Drought Severity Index for central New Mexico (1895-2002). Note the periods of low (dry) indices during the periods of 1898-1905 and 1950-1957.

New Mexico - Division 05: 189S-2002 (Monthly Averages)

Figure 15.4 Palmer Drought Severity Index for central New Mexico (1895-2002). Note the periods of low (dry) indices during the periods of 1898-1905 and 1950-1957.

for massive die-offs of livestock and shrub encroachment on southwestern rangelands in the 1890s. The period 1950-1957 is known as the "50's drought." In the Southwest, this was much more severe than the "Dust Bowl" drought of the 1930s in the panhandle region of Oklahoma and Texas. Table 15.2 lists the years with mean annual PDSI values less than minus 2.0. These two dry periods indicated by the PDSI are consistent with the tree-ring chronology and reflect the low rainfall periods in the Socorro precipitation record.

The PDSI values are significantly related to the reconstructed precipitation depths. A regression analysis using reconstructed precipitation from the tree-rings (Betancourt et al. 1993; Parks et al. in press) and the mean annual PDSI values from Climate Region 5 (Middle Rio Grande Valley) produced an R2 = 0.46 (p < 0.0001, n = 95; figure 15.5). Among months, May and June exhibited the strongest relations (R2 = 0.497, p < 0.0001, n = 95 and R2 = 0.469, p < 0.001, n = 95), respectively. Thus, the reconstructed precipitation depths were statistically related to regional moisture conditions. However, the considerable unexplained variation, possibly due to soil, hydraulic, and orographic conditions at the Arroyo de Milagro site, cautions against unbridled application of the record to the region.

Table 15.2 Years from 1895 to 2000 with mean annual Palmer Drought Severity Index (PDSI) of minus 2.0 or less. Note multiyear periods of drought in early 1900s and 1950s.



















































Influence of Climate on Sevilleta's Biome Transition Zone

The Sevilleta LTER site is at the interface of several major biomes, including short grass prairie, Chihuahan Desert shrubland, and pinon-juniper woodland. Combined effects of potential evapotranspiration and available soil moisture appear to regulate the distributions of the various biomes. For example, where uniform soils occur over locally rough terrain, nuances of slope and aspect regulate the surface energy supply and produce repeating vegetation patterns (figure 15.6). Desert shrubs such as Larrea tridentata dominate on south-facing slopes with high moisture deficits. A few meters away, Juniperus monosperma dominates at the same elevation on neighboring north-facing slopes. Inspections of aerial photographs taken before and after the 1950s drought indicate that junipers found at intermediate east and west facing slopes died where conditions became marginal during the drought. Today, the two dominant species meet in such vicinities. Repeat photography indicates that L. tri-dentata became more widely established after the 1950s drought (figure 15.7), apparently reflecting the massive expansion of desert shrubland in New Mexico (Grover and Musick 1990).

We used preliminary measurements of the relative growth rates of L. tridentata and J. monosperma as functions of annual precipitation to examine the hypothesis that the species were capable of growing in the distant past at Arroyo de Milagro. The periodic wet and dry epochs over the last 400 years (figure 15.3c) should have created opportunities for species such as L. tridentata to establish. However, satisfactory resource availability would not overcome dispersal limitations or competition that can also affect establishment. Ongoing field experiments address the competitive interactions of these species.

Longitudinal studies of 10 marked twigs from each of 20 shrubs of each species at 7 sites (200 <PPT< 340 mm) were used to calibrate second-order polynomial regressions of mass specific growth rate as a function of annual precipitation (B. T. Milne, C. Restrepo, and W. Pockman, unpubl. data, 2001). The mean relative

Figure 15.5 Regression of the Palmer Drought Severity Index (PDSI) values and the precipitation record reconstructed from the Arroyo de Milagro tree-ring data.

change in L. tridentata biomass (B) was dB/Bdt = -137.2 + 1.28 PPT - 0.0029 PPT2, where PPT is annual precipitation (mm) in 2000. For J. monosperma, dB/Bdt = -6.83 + 0.05 PPT-0.0001 PPT2. We simulated annual net primary productivity for the years 1598-1991, assuming that plants began with 1 kg/m2 of biomass (i.e., were not limited by dispersal) and had the potential to grow logistically to a carrying capacity of 50 kg/m2. We assumed no competition. Thus, the simulation estimated the potential net primary productivity for each species given the series of annual precipitation depths estimated from the tree-ring record.

The relatively high mean annual precipitation at Arroyo de Milagro favored the growth of J. monosperma more than L. tridentata. Several decades showed persistent epochs of high juniper productivity that coincided with lethal conditions for L. tridentata (figure 15.8). Suitable conditions for L. tridentata appeared throughout the record but were interrupted repeatedly by periods of no growth. We concluded that climatic variability, dispersal limitation, and competition are possible explanations for the lack of L. tridentata in the area until recent times.

Of course, variability across the broader landscape could maintain source populations for either species. Source populations would ameliorate the effects of climate variability at a given location by providing an infusion of propagules. Thus, a comprehensive view would include a metapopulation approach to understand the role of temporal variation through space (Keymer et al. 2000). Climate fluctuations can shift the edges of populations as species advance beyond their previous distribution or are eliminated from occupied areas. Similarly, at the edge of a species distribution, variation in microclimate may create locally patchy distributions (Holt and Keitt 2000).

Figure 15.6 View of a small watershed at the Sevilleta LTER site, showing juniper savanna vegetation on north-facing slope (left) contiguous with Chihuahuan Desert vegetation on the south-facing slope (right).

One example of a physiological limit that is particularly relevant in arid ecosystems derives from the plant's requirement for water to sustain transpiration. Water transport through the xylem is subject to interruption by cavitation caused either by high xylem tensions associated with drought or by cycles of freezing and thawing in the xylem (Tyree and Sperry 1989). The drought or freezing conditions required to cause cavitation vary considerably among species and are determined by structural features of the xylem, pit membrane pore diameter in the case of drought (Sperry and Tyree 1988) and xylem conduit volume in the case of freezing (Davis et al. 1999; Sperry et al. 1994). This quantifiable link to xylem structure makes the definition of a physical limit to plant function more straightforward than for some physiological parameters for which the functional limit is difficult to predict. Although some species can repair the effect of cavitation (Holbrook and Zwieniecki 1999), hydraulic limits are largely fixed in an individual, making the maintenance of water transport in the individuals dependent on the drought and freezing conditions they experience while active.

Drought Effects

Highly variable precipitation reduces the chance that soil water conditions that favor germination and establishment will occur in the necessary sequence (Neilson 1986). Although dominant woody species at Sevilleta LTER (e.g., Pinus edulis, J. monosperma, L. tridentata) are highly drought tolerant (Linton et al. 1998; Pock-

Figure 15.7 Photographs of Palo Duro Canyon area, Sevilleta LTER site. Top: Cattle herd near windmill drinking area, circa 1928. Bottom: Same view in 1998. Note the invasion by creosotebush (Larrea tridentata) and grasses (Sporobolus spp.).
Figure 15.8 Simulated net primary productivity of Juniperus monosperma (solid thin line) and Larrea tridentata (filled dots) with respect to cumulative precipitation over the 394-year surrogate precipitation history from Arroyo de Milagro, Sevilleta LTER.

man and Sperry 2000), the limited extent of seedling root systems exposes them to extremely dry shallow soil. Under such conditions, extensive cavitation leads to mortality (Williams et al. 1997). Once established, the dominant shrub species approach the point of complete cavitation only rarely (Pockman and Sperry 2000). Conditions during extended drought periods such as the 1950s drought in New Mexico (Allen and Breshears 1998) are likely to approach the physiological limits of these species to drought induced cavitation.

Freezing Effects

At Sevilleta, the transition from shrubland, dominated by L. tridentata, to grassland in the north represents the northern limit of the continuous distribution of L. tri-dentata. Although its evergreen habit enables L. tridentata to exploit favorable conditions at any time, it must also maintain water transport during freezing conditions to support its evergreen foliage. Historical reports of L. tridentata at its northern limit indicate that extreme freezing events lead to heavy dieback of aboveground growth (Cottam 1937). Stems of mature L. tridentata in southern Arizona exhibit no embolism following freezing between 0° and -10° C, after which embolism increases linearly with decreasing temperature until embolism is complete at -16° to -20° C (Pockman and Sperry 1997). A coarse-scale analysis of long-term climate data showed that this critical temperature range corresponds to the northern limit of the species in the Sonoran and Mojave Deserts but fails to account for the north ern extension of the range in the Rio Grande valley to its limit at Sevilleta. Recent data suggest that L. tridentata at Sevilleta are more resistant to freezing-induced cavitation than those in Arizona (Martinez-Vilalta and Pockman, 2002). Biogeo-graphic variation provides an opportunity to examine the role of this physiological limit in determining the distribution of the species.

Ecosystem Feedbacks on Sevilleta's Climate

Although it is clear that temperature and moisture influence the distributions of plants at the Sevilleta LTER site, there is some evidence that the vegetation patterns have feedbacks on the local climate as well. One example of this occurs in an area of the Sevilleta that is dominated by L. tridentata (figure 15.7). It is widely hypothesized that anthropogenic disturbances of arid lands in Mexico and the American Southwest, in concert with drought cycles (e.g., the great droughts of the 1890s and 1950s; Hastings and Turner 1965), have facilitated the recent acceleration of the range expansion of woody shrubs. Rangeland overgrazing by excessive cattle, sheep and horses in the late nineteenth century, coupled with extended droughts, favored desert shrub (L. tridentata and mesquite, Prosopis spp.). Bray (1901, p. 289) noted the speed of the invasions at the turn of the century in west Texas: "Regarding the establishment of woody vegetation, it is the unanimous testimony of men of long observation that most of the chaparral [=L. tridentata] and mesquite covered country was formerly open grass prairie." Encroachment of desert shrubs northward and into grasslands continued during the twentieth century, particularly in New Mexico and west Texas (Gardner 1951; Branscomb 1958; Humphrey 1958; Buffington and Herbel 1965; York and Dick-Peddie 1969; Hen-nessy et al. 1983; Humphrey 1987; McPherson et al. 1988; Grover and Musick 1990).

Invasions of shrub species into grasslands not only have considerable consequences for ecosystem dynamics (Schlesinger et al. 1990) and human activities (Mainguet 1994), but also for potential alterations of mesoscale climatic conditions. Hayden (1998) reviewed the existing literature and analyzed large-scale temperature patterns, relating the observed minimum temperatures in deserts of the American Southwest to the predicted dew-point temperatures of the same sites. Winter minimum temperatures ranged up to 8°C warmer than predicted (Hayden 1998); at the Sevilleta LTER site, winter minimum temperatures could be as much as 4°C warmer than predicted from dew-point temperatures. The reason for this difference was attributed to nonmethane hydrocarbons released from desert vegetation (particularly terpenes from desert shrubs like L. tridentata); these hydrocarbons functioned as "greenhouse gasses" and decreased the emissivity of the local atmosphere above the vegetation, preventing heat loss and raising nighttime minima.

At the Sevilleta LTER site, two meteorological stations are located within 5 km of each other at the same elevation on McKenzie Flats; one is in a community dominated by L. tridentata, whereas the other is surrounded by grassland. The station in creosotebush vegetation typically registers higher nighttime minimum tempera-

Temperature °C

25 20 15

25 20 15

1 23456789 10

Day of the Year Station 40 = Grassland 49 = Shrubs - 40 49

1 23456789 10

Day of the Year Station 40 = Grassland 49 = Shrubs - 40 49

Figure 15.9 Air temperatures above desert shrub vegetation and grassland vegetation at the Sevilleta LTER site, January 1-10, 2000; weather stations are 4.4 km apart. Note warmer minimum temperatures above desert shrub vegetation.

tures than the station in the grassland (figure 15.9), particularly during calm periods (high winds mix the atmosphere over both sites). Atmospheric warming reduces the probability of killing-frosts, thereby increasing the likelihood of survival and population increase of these desert shrub species. Thus, it may be that desertification of grasslands by woody shrub invasions is accelerated by the invading shrubs via changes in atmospheric chemistry that ultimately alter the temperature regime, allowing greater survivorship and expansion of the shrub populations by frost reduction. Future research on this hypothesis will be needed to determine whether, and by how much, vegetation changes can influence the regional climates of arid and semiarid ecosystems.


Ecologists have been slow to measure or model biological responses to longer periodicities in climate. Most gap models, for example, simulate climate variability as randomized variance around an annual mean. Inserting decadal climatic oscillations in gap models, however, has been shown to either induce periodicity or shift the mean in biomass (Yeakley et al. 1994). A general characterization of the Sevil-leta climate includes seasonal, interannual, and decadal scales of variation.

The 41- to 63-year periodicity in annual precipitation constitutes a slowly changing climatic context within which higher frequency ecological responses occur. Alterations of wild populations by management, disease, or predation might result in different responses, depending on whether the oscillation is on an upswing or a downswing. Efforts at reforestation or land reclamation would probably be most successful in the wetting phase of the cycle, during which a 20- to 30-year period of relatively wet weather would benefit the efforts. Attempts to maintain marginally productive rangeland operations during the ensuing dry phase of the oscillation are at risk of failure. Indeed, it may be no accident that the Sevilleta National Wildlife Refuge was created from a working ranch only 10 years after the end of the 1950s drought when the CPA was at an all-time low (figure 15.3c).

Policy makers and financial institutions might anticipate that economic collapses follow droughts that stimulate abrupt changes in land ownership. The Sevilleta tree-ring record, limited to only 7 multiyear droughts of different severities and durations, suggests a 52-year recurrence interval for major droughts (figure 15.3b), with the last drought period persisting between 1942 and 1972. We presently lack the necessary time depth, regional coverage, and understanding of the global-scale ocean-atmosphere interactions that underlie this apparent periodicity. We also recognize that these periodicities may be inherently unstable and could be modulated by various phenomena, including anthropogenic forcing of climate. Hence, we exercise caution in making any prediction about imminent drought in south-central New Mexico. Nevertheless, we point out that extreme drought in the early part of the twenty-first century is within the realm of expectation and that human population growth and demand for water and other resources in south-central New Mexico has amplified sensitivity to drought.

Land managers might reflect on the experiences of the 1890s. The introduction of railroads created an incentive to stock the range because of easier access to markets. Simultaneously, introduction of windmills enabled ranchers to supply drinking water for livestock during the drought. However, without precipitation, forage plants were unable to withstand high stocking rates, leading to long-lasting depletion of soil organic matter and seed banks. In future droughts, apparent technological solutions such as precision farming methods, optimized irrigation strategies, and genetically engineered crops or livestock may actually increase the risk to various components of the landscape or ecosystem, just as wells and trains did in the past. The ecology of semiarid lands is subject to environmental variation on the order of a human career and thus should probably be managed at timescales of several human generations.


Variation in precipitation occurs at seasonal, interannual, and decadal timescales. Ecological and economic consequences related to land use and resource management drive the search for repeatable patterns ascertained from direct and surrogate climate records. We studied a 394-year record (1598-1991) of precipitation derived from the annual rings of piñon pine (Pinus edulis) at the Sevilleta National Wildlife

Refuge and Long-Term Ecological Research site in central New Mexico, United States. A significant 52-year periodicity (standard deviation 11.3 years) of precipitation coincided with the major regional droughts of the 1890s and 1950s. Long-term ecological consequences of decade-long droughts pertain to the establishment of novel species, physiological stress, feedbacks between plants and the atmosphere, and economic repercussions related to land use. Simulated net primary productivity of Juniperus monosperma (one-seeded juniper) and Larrea tridentata (creosotebush) indicate that suitable conditions for L. tridentata growth occurred intermittently during the last four centuries. Assessments of the droughts of the 1890s and 1950s suggest that future technological attempts to ameliorate the effects of drought should minimize unforeseen consequences for various components of the landscape. Occasional decadal oscillations in annual precipitation are a major ecological factor in the region.

Acknowledgments Dave Meko, Mike Dettinger, and two anonymous reviewers guided the final outcome of this effort. Support was provided by NSF grants DEB 9910123 to BTM and DEB 0080529. Sevilleta LTER publication no. 263.


Allen, C., and D. Breshears. 1998. Drought-induced shift of a forest-woodland ecotone: Rapid landscape response to climate variation. Proceedings of the National Academy USA 95:14839-14842.

Andrade, E.R., and W.D. Sellers. 1988. El Niño and its effect on precipitation in Arizona. Journal of Climatology 8:403-410.

Archer, S., C. Scrifres, C.R. Bassham, and R. Maggio. 1988. Autogenic succession in a subtropical savanna: Conversion of grassland to thorn woodland. Ecological Monographs 58:111-127.

Betancourt, J.L., E.A. Pierson, K.A. Rylander, J.A. Parks, and J.S. Dean. 1993. Influence of History and Climate on New Mexico Pinyon-Juniper Woodlands. Pages 53-62 in Managing Juniper Ecosystems for Sustainablity and Social Needs. USDA Forest Service General Technical Report RM-236. Rocky Mountain Forest and Range Experiment Station. U.S. Department of Agriculture. Fort Collins, Colorado.

Biondi, F., A. Gershunov, and D.R. Cayan. 2001. North Pacific decadal climate variability since 1661. Journal of Climate 14:5-10.

Branscomb, B. L. 1958. Shrub invasion of a southern New Mexico desert grassland range. Journal of Range Management 11:129-132.

Bray, W.L. 1901. The ecological relations of the vegetation of western Texas. Botanical Gazette 32:99-123, 195-217, 262-291.

Brown, D.E. 1982. Biotic Communities of the American Southwest—United States and Mexico. Desert Plants 4:1-342.

Buffington, L.C., and C.H. Herbel. 1965. Vegetational changes on a semidesert grassland range from 1858 to 1963. Ecological Monographs 35:139-164.

Cayan, D.R., and R.H. Webb. 1992. El Niño/Southern Oscillation and streamflow in the Western United States. Pages 29-68 in H.F. Diaz and V. Markgraf, editors, El Niño Historical and Paleoclimatic Aspects of the Southern Oscillation. Cambridge University Press, Cambridge.

Cayan, D.R., M.D. Dettinger, H.F. Diaz, and N.E. Graham. 1998. Decadal variability of precipitation over western North America. Journal of Climate 11:3148-3166.

Cottam, W.P. 1937. Has Utah lost claim to the lower Sonoran zone? Science 85:563-564.

Davis, S., J. Sperry, and U. Hacke. 1999. The relationship between xylem conduit diameter and cavitation caused by freezing. American Journal of Botany 86:1367-1372.

Delworth, T., S. Manabe, and R.J. Stouffer. 1993. Interdecadal variations of the thermohaline circulation in a coupled ocean-atmosphere model. Journal of Climate 6:19932011.

Dettinger, M.D., M. Ghil, C.M. Strong, W. Weibel, and P. Yiou. 1995. Software expedites singular-spectrum analysis of noisy time series, Eos, Trans. American Geophysical Union 76(2):12, 14, 21.

Dettinger, M.D., D.R. Cayan, G.J. McCabe, Jr. and K.T. Redmond. 2000. Winter-spring 2001United States streamflow probabilities based on anticipated neutral ENSO conditions and recent NPO status. Experimental Long-Lead Forecast Bulletin 9:55-60.

Dettinger, M.D., D.S. Battisti, R.D. Garreaud, G.J. McCabe, and C.M. Bitz. 2001. Inter-hemispheric effects of interannual and decadal ENSO-like climate variations on the Americas. Pages 1-16 in V. Markgraf, editor, Present and Past Interhemispheric Climate Linkages in the Americas and their Societal Effects. Academic Press, San Diego, California.

Enfield, D.B., and A.M. Mestas-Nuñez. 1999. Multiscale variabilities in global sea surface temperatures and their relationships with tropospheric climate patters. Journal of Cli-mate12:2719-2733.

Feller, W. 1968. An introduction to probability theory and its applications. Vol. 1. Wiley, New York.

Gardner, J.L. 1951. Vegetation of the creosotebush area of the Rio Grande valley in New Mexico. Ecological Monographs 21:379-403.

Gedalof, Z., and D.J. Smith. 2001. Interdecadal climate variability and regime-scale shifts in Pacific North America. Geophysical Research Letters 28:1515-1518.

Gerschnov, A., and T.P. Barnett. 1998. Interdecadal modulation of ENSO teleconnections. Bulletin of the American Meteorological Society 79:2715-2725.

Grissino-Mayer, H.D., C.H. Baisan, and T.W. Swetnam. 1997. A 1,374-year reconstruction of annual precipitation for the southern Rio Grande Basin. Final Report, Department of Defense, Legacy Program, Ft. Bliss, Texas.

Grover, H.D., and H.B. Musick. 1990. Shrubland encroachment in southern New Mexico, U.S.A.: An analysis of desertification processes in the American Southwest. Climatic Change 17:305-330.

Hare, S.R., S. Minobe, W.S. Wooster, and S. McKinnell. 2000. An introduction to the PICES symposium on the nature and impacts of North Pacific climate regime shifts. Progress in Oceanography 47(2-4):99-102.

Harrington, J.A., R.S. Cerveny, and R.C. Balling, Jr. 1992. Impact of the Southern Oscillation on the North American southwest monsoon. Physical Geography 13:318-330.

Hastings, J.R., and R.M. Turner. 1965. The changing mile. An ecological study of vegetation change with time in the lower mile of an arid and semiarid region. University of Arizona Press, Tucson.

Hayden, B.P. 1998. Ecosystem feedbacks on climate at the landscape scale. Philosophical Transactions of the Royal Society London B, 353:5-18.

Hennessy, J.T., R.P. Gibbens, J.M. Tromble, and M. Cardenas. 1983. Vegetation changes from 1935 to 1980 in mesquite dunelands and former grasslands of southern New Mexico. Journal of Range Management 36:370-374.

Herbel, C.H., F.N. Ares, and R.A. Wright. 1972. Drought effects on a grassland range. Ecology 53:1084-1094.

Higgins, R.W., and W. Shi. 2000. Dominant factors responsible for interannual variability of the summer monsoon in the southwestern United States. Journal of Climate 13(4):759-776.

Holbrook, N. M., and M.A. Zwieniecki. 1999. Embolism repair and xylem tension: Do we need a miracle? Plant Physiology 120:7-10.

Holt, R.D., and T.H. Keitt. 2000. Alternative causes for range limits: A metapopulation perspective. Ecology Letters 3:41-47.

Humphrey, R.R. 1958. The desert grassland. A history of vegetational change and an analysis of causes. Botanical Review 24:193-251.

Humphrey, R.R. 1987. 90 years and 535 miles: Vegetation changes along the Mexican border. University of New Mexico Press, Albuquerque.

Kahya, E., and J.A. Dracup. 1993. U.S. streamflow patterns in relation to the El Niño/Southern Oscillation. Water Resources Research 29:2491-2503.

Keymer, J. E., P.A. Marquet, J.X. Velasco-Hernandez, and S.A. Levin. 2000. Extinction thresholds and metapopulation persistence in dynamic landscapes. American Naturalist 156:478-494.

Latif, M., and T.P. Barnett. 1996. Decadal climate variability over the North Pacific and North America: Dynamics and predictability. Journal of Climate 9:2407-2423.

Linton, M.J., J.S. Sperry, and D.G. Williams. 1998. Limits to water transport in Juniperus os-teosperma and Pinus edulis: Implications for drought tolerance and regulation of transpiration. Functional Ecology 12:906-911.

Mainguet, M. 1994. Desertification: Natural background and human mismanagement. Springer-Verlag, New York. 314 pp.

Mantua, N.J., and S.R. Hare. 2002. The Pacific Decadal Oscillation. Journal of Oceanography 58:35-44.

Martínez-Vilalta, J., and W.T. Pockman. 2002. The vulnerability to freezing-induced xylem cavitation of Larrea tridentate (Zygophyllaceae) in the Chihuahuan desert. American Journal of Botany 89:1916-1924.

McCabe, G.J., and M.D. Detinger. 1999. Decadal variations in the strength of ENSO tele-connections with precipitation in the western United States. International Journal of Climatology 19:1069-1079.

McPherson, G.R., H.A. Wright, and D.B. Wester. 1988. Patterns of shrub invasion in semiarid Texas grasslands. American Midland Naturalist 120:391-397.

Minobe, S. 1997. A 50-70 year climatic oscillation over the North Pacific and North America. Geophysical Research Letters 24:683-686.

Minobe, S. 1999. Resonance in bidecadal and pentadecadal climate oscillations over the North Pacific: Role in climatic regime shifts. Geophysical Research Letters 26:855858.

Mitchell, V.L. 1976. The regionalization of climate in the western United States. Journal of Applied Meteorology 15:920-927.

Molles, M.C., Jr., and C.N. Dahm. 1990. A perspective on El Niño and La Niña: Global implications for stream ecology. Journal of the North American Benthological Society 9:68-76.

Molles, M.C., Jr., C.N. Dahm, and M.T. Crocker. 1992. Climatic variability and streams and rivers in semi-arid regions. Pages 197-202 in R.D. Robarts and M.L. Bothwell, editors, Aquatic ecosystems in semi-arid regions: Implications for resource management. Environment Canada, Saskatoon, Canada.

Neilson, R.P. 1986. High resolution climatic analysis and southwest biogeography. Science 232:27-34.

Palmer W.C. 1965 Meteorological drought. Research Paper Number 45, U.S. Department of Commerce Weather Bureau. Washington, D.C.

Parks, J.A., J.S. Dean, and J.L. Betancourt, in press, Tree rings, drought and the Pueblo Abandonment of south-central New Mexico in the 1670s. In D.E. Doyel and J.S. Dean, editors, Environmental Change and Human Adaptation in the Ancient Southwest. University of Utah Press, Salt Lake City, Utah.

Peng, C.S., and N. Buras. 2000. Practical estimation of inflows into multireservoir system. Journal of Water Resources Planning and Management-ASCE 126:331-334.

Pockman, W.T., and J.S. Sperry. 1997. Freezing-induced xylem cavitation and the northern limit of Larrea tridentata. Oecologia 109:19-27.

Pockman, W.T., and J.S. Sperry. 2000. Vulnerability to cavitation and the distribution of Sonoran Desert vegetation. American Journal of Botany 87:1287-1299.

Redmond, K.T., and R.W. Koch. 1991. Surface climate and streamflow variability in the western United States and their relationship to large-scale circulation indices. Water Resources Research 27:2381-2399.

Regensberg, A. 1996. General dynamics of drought, ranching and politics in New Mexico, 1953-1961. New Mexico Historical Review, January:25-49.

Schlesinger, W.H., J.F. Reynolds, G.L. Cunningham, L.F. Huenneke, W.M. Jarrell, R.A. Virginia, and W.G. Whitford. 1990. Biological feedbacks in global desertification Science 247:1043-1048.

Schneider, N., and A.J. Miller. 2001. Predicting western North Pacific Ocean climate. Journal of Climate 14:3997-4002.

Sperry, J.S., and M.T. Tyree. 1988. Mechanism of water stress-induced xylem embolism. Plant Physiology 88:581-587.

Sperry, J.S., K.L. Nichols, J.E.M. Sullivan, and S.E. Eastlack. 1994. Xylem embolism in ring-porous, diffuse-porous and coniferous trees of northern Utah and interior Alaska. Ecology 75:1736-1752.

Swetnam, T.W., and J.L. Betancourt. 1990. Fire-Southern Oscillation relations in the southwestern United States. Science 249:1017-1020.

Swetnam, T.W., and J.L. Betancourt. 1998. Mesoscale disturbance and ecological response to decadal climatic variability in the American Southwest. Journal of Climate 11:31283147.

Thomas, H.E. 1963. General summary of the effects of the drought in the Southwest: Drought in the Southwest, 1942-1956. U.S. Geological Survey Professional Paper 372-H.

Tyree, M.T., and J.S. Sperry. 1989. Vulnerability of xylem to cavitation and embolism. Annual Review of Plant Physiology and Molecular Biology 40:19-38.

Valentine, K.A., and J.J. Norris. 1964. A comparative study of soils of selected creosotebush sites in southern New Mexico. Journal of Range Management 17:23-32.

Villalba, R., R.D. D'Arrigo, E.R. Cook, G.C. Jacoby, and G. Wiles. 2001. Decadal-scale climate variability along the extratropical western coast of the Americas: Evidence from tree-ring records. Pages 155-172 in Vera Markgraf, editor, Interhemispheric Climate Linkages. Academic Press, San Diego.

Williams, J.E., S.D. Davis, and K. Portwood. 1997. Xylem embolism in seedlings and re-sprouts of Adenostoma fasciculatum after fire. Australian Journal of Botany 45:291300.

Yeakley, J.A., R.A. Moen, D.D. Breshears, and M.K. Nungesser. 1994. Response of North

American ecosystem models to multi-annual periodicities in temperature and precipitation Landscape Ecology 9:249-260. York, J.C., and W.A. Dick-Peddie. 1969. Vegetation changes in southern New Mexico during the past hundred years. Pages 155-166 in W.G. McGinnies and B.J. Goldman, editors, Arid Lands in Perspective. University of Arizona Press, Tucson.

Was this article helpful?

0 0

Post a comment