Aboveground ANPP responses to natural precipitation
The responses of ANPP of our three dominant perennial species (D. leiophyllum, a shrub; O. phaeacantha, a succulent; and B. curtipendula, a grass) to variation in the timing and magnitude of natural precipitation (annual and seasonal) during the 5-yr study period varied for each species. Dasylirion leiophyllum exhibited its highest ANPP during the wettest year (2004) when precipitation was 55% above average, precipitation events were 44% above average, and large precipitation events (> 20 mm) were 140% more frequent than average. These frequent large precipitation events, combined with frequent smaller events, wet the upper and lower soil layers for long periods of time, allowing roots in both zones to utilize soil moisture for most of the growing season, thereby promoting plant growth (Gibbens & Lenz, 2001). Apparently, frequent large precipitation events in the wet year, in which these events were twice as frequent as in any other year, were key determinants of productivity in this deeper rooted shrub. Fravolini et al. (2005) found similar results with another deep-rooted shrub, mesquite (Prosopis velutina), where large rain events during wet summers resulted in increased water uptake and photosynthesis, especially in course-textured soils, leading to increased biomass.
In B. curtipendula, ANPP was highest in an average precipitation year following a dry year, which suggests that total annual precipitation was not a major determinant of productivity in this species. The winter precipitation was below average, but summer precipitation was average, with most of the precipitation occurring in small (< 5 mm) events during this period of high physiological activity. During the summer, there were also very few inter-pulse periods > 10 d, indicating that the soil was rarely dry for extended periods of time. These data suggest that frequent, small precipitation events with relatively few extended dry inter-pulse periods promote productivity in this shallow-rooted bunchgrass. Similarly, Jobbagy & Sala (2000) observed a weak correlation of ANPP with annual precipitation in several grass species in the Patagonian steppe, as ANPP was more strongly correlated with seasonal precipitation amounts and temperature. This suggests that grass ANPP was primarily responsive to the seasonal timing and magnitude of precipitation and subsequent soil moisture, rather than to the amount of precipitation received annually. In our study, ANPP in B. curtipendula may be more influenced by seasonal precipitation patterns, mainly summer precipitation with frequent precipitation events and few long inter-pulse periods, rather than total summer precipitation amounts. In addition, adequate winter and spring precipitation may maintain root development, allowing these plants to fully exploit water availability when physiologically active during the late spring and summer months (Bates et al., 2006; Muldavin et al., 2008). Furthermore, the decline of ANPP during wetter years (2003–2004) may be attributable to other limiting factors (e.g. NH4-N and NO3-N) or increased competition as a result of greater plant density (Yahdjian & Sala, 2006; Muldavin et al., 2008).
In O. phaeacantha, ANPP was highest in an average precipitation year (2003), when plants received average winter precipitation, as well as average fall precipitation in the previous year. This precipitation pattern differed from that in other study years, when plants were exposed to either dry winters, or both a dry winter and a dry fall in the previous year. In the spring of 2003, precipitation magnitude was below average, but most precipitation events were small (< 5 mm) with most inter-pulse periods < 20 d, indicating that soil moisture was probably adequate at shallow rooting depths (5–10 cm), where O. phaeacantha roots are most abundant and can readily utilize water (Dougherty et al., 1996). Pad production occurs in mid/late spring (Powell, 1998) and is largely dependent on prior fall and winter precipitation, when water is stored in the soil and in the succulent pads of O. phaeacantha, rather than only current spring precipitation events (Muldavin et al., 2008).
Aboveground ANPP response to supplemental precipitation
ANPP responses to increased supplemental seasonal precipitation also varied depending on the species. Supplemental seasonal precipitation did not influence ANPP in D. leiophyllum, but greater ANPP was observed for B. curtipendula. In B. curtipendula, supplemental winter precipitation in an average year (2002), following a very dry year, generated a large positive ANPP response in our winter watering treatment. The large pulse of supplemental water (e.g. 25 mm, which constituted more than half of the natural winter precipitation) plus the very dry conditions preceding the supplemental winter precipitation event was sufficient to initiate plant growth when temperatures increased in the spring. The significant impact of this large single rainfall event indicates the critical importance of winter precipitation in this grass species. In subsequent years, winter precipitation was generally average and the impact of the W treatment was no longer observed. The SW treatments in 2002 did not generate the same response as the W treatment, suggesting that summer additions altered the impact of the winter precipitation treatment. Bates et al. (2006) observed a similar response during a 7-yr study in the northern Great Basin where shallow-rooted grasses produced greater biomass when the majority of the precipitation was shifted from spring to winter. Furthermore, it is possible that the upper soils may have approached water-holding capacity during the summers of 2003 and 2004, resulting in the treatments being less effective in triggering individual growth responses (Muldavin et al., 2008).
In O. phaeacantha, SW treatments increased ANPP following two consecutive years of average precipitation. Pad production in O. phaeacantha depends largely on water that is available before mid-spring, when new pads are produced. Therefore, supplemental water in 2002 for the SW treatment in the summer during an average summer and fall rainfall period and in the winter during an average winter followed by a dry spring may have delivered sufficient additional water for increased pad production in 2003. However, because O. phaeacantha pads store water, it is unclear if increased winter precipitation could also significantly contribute to increased pad production the following summer.
Best predictors of ANPP
When we further explored ANPP responses to other key environmental and soil parameters to determine the potential impact of these variables on the ANPP of each species, we found that the responses varied for each species and between sampling years. For D. leiophyllum, soil organic matter (RDA analysis) and soil NH4-N (Kendall tau correlation) may have strong impacts on aboveground ANPP during wet years. However, during drier years, ANPP was mainly affected by climatic variables (e.g. small precipitation events and shorter inter-pulse periods) rather than soil variables (e.g. NH4-N and NO3-N). The importance of the woody caudex of this species to plant growth dynamics has never been fully investigated, but it may allow the plant to store sufficient quantities of water and nutrients in wetter years. The ability to store water in some shrubs allows survival through long drought periods (Barker et al., 2006). Dasylirion leiophyllum is a very long-lived and slow-growing plant, potentially establishing an ‘island of fertility’ which may provide the plant with a localized nutrient supply when soil moisture is not limiting (Schlesinger & Pilmanis, 1998; Reynolds et al., 1999). Dasylirion leiophyllum also had roots within the upper soil horizons, allowing it to compete with grasses for soil moisture, as well as having deeper roots giving it access to water in lower soil horizons (Scott et al., 2000; Gibbens & Lenz, 2001). Because of this extensive root system, it is difficult to clearly distinguish which climatic variables have a greater impact on ANPP as indicated by the Kendall tau correlation.
For B. curtipendula, soil organic matter (RDA analysis) and NO3-N (Kendall tau correlation) had strong impacts on ANPP in an average precipitation year, suggesting that nitrogen mineralization rates may be a significant regulator of ANPP when soil moisture is sufficient for growth. In above-average precipitation years, soil NH4-N and NO3-N concentrations had stronger impacts on ANPP than in average rainfall years (RDA analysis), perhaps as a result of changes in available nitrogen. During a dry year, climatic variables and soil NH4-N had a greater impact on ANPP because soil moisture was apparently scarcer in the upper soil horizons. It is possible that seasonal precipitation, in particular winter precipitation, and longer inter-pulse duration may have a greater impact on ANPP in B. curtipendula, especially during a dry season when soil moisture is less available (Bates et al., 2006; Yahdjian & Sala, 2006; Muldavin et al., 2008).
ANPP in O. phaecantha appears to be primarily regulated by climatic variables rather than soil variables, particularly precipitation magnitude and length of inter-pulse periods (RDA analysis and Kendall tau correlation). Following an average precipitation year, ANPP may be affected more strongly by shorter inter-pulse period duration (11–20 d) than by annual precipitation; however, when annual precipitation was above-average, both precipitation magnitude and the frequency of shorter inter-pulse periods had a greater impact on ANPP. An exception to this may occur with winter precipitation (e.g. when there was greater winter precipitation in 2003 and 2004, but ANPP in O. phaeacantha still declined in 2004). This may be a result of changes in soil NO3-N, because ANPP in O. phaeacantha was negatively correlated with NO3-N concentrations, suggesting that increased nitrogen availability may limit ANPP (Whitford, 1986; Austin et al., 2004; Havstad et al., 2006).
It is possible that there may be a memory or lag effect caused by past precipitation events as a result of pad water storage in O. phaeacantha (Dougherty et al., 1996; Schwinning et al., 2004). During a dry year, soil pH, small precipitation magnitude events, and inter-pulse periods of 11–20 d had significant impacts on ANPP in O. phaeacantha. There was almost no new pad production during the dry spring of a dry year (2006). During dry winters and springs, O. phaeacantha may maintain current pads rather than promote vegetative and sexual reproduction, resulting in very little detectable change in ANPP (Powell & Weedin, 2004). In the RDA analysis, O. phaeacantha ANPP was positively correlated with inter-pulse periods of 11–20 d and negatively related to inter-pulse periods > 20 d for all years except 2004, which experienced more frequent shorter inter-pulse periods. Therefore, medium inter-pulse periods are apparently greater regulators of ANPP in O. phaeacantha than precipitation magnitude and amount.
In this Chihuahuan Desert grassland ecosystem in Big Bend National Park, ANPP is limited not only by soil moisture and temperature constraints, but also by soil NO3-N and NH4-N concentrations. During consecutive years of average and above-average precipitation, extractable nitrogen pools were negatively correlated with annual precipitation, intermediate magnitude events, and shorter inter-pulse periods (RDA analysis). This may suggest that soil N is assimilated as soil moisture becomes available via precipitation events (Bell et al., 2008). However, during an average rainfall year following a dry year (2002) or a dry year following an average rainfall year (2006), extractable nitrogen pools were positively correlated with annual precipitation, intermediate magnitude events, and shorter inter-pulse periods (Muldavin et al., 2008). This may indicate a seasonal soil N build-up during seasons with sporadic precipitation. Furthermore, as soil moisture becomes available via successive precipitation events, soil nitrogen becomes soluble and readily available for plant and microbial uptake.
Soil nitrogen is commonly limiting in desert grasslands, especially in wet years as a result of declines in nitrogen availability and immobilization from previous year ANPP (Whitford, 1986; Austin et al., 2004; Havstad et al., 2006). Wind and water erosion, especially in sites of low plant cover, may also cause shifts in available nitrogen in wetter years (Schlesinger & Pilmanis, 1998; Havstad et al., 2006). Plant cover in the sotol grassland site is c. 50%, resulting in a patchy landscape. Consequently, the bare soil-patches could experience soil nitrogen loss through runoff during wetter years, which may cause vegetation shifts in arid ecosystems as a result of limited N availability (Schlesinger et al., 2000; Muldavin et al., 2008).
Many studies have shown that ANPP increases with greater annual precipitation. In this sotol grassland site in the Chihuahuan Desert, there was no universal predictor of ANPP, as the response of each species to precipitation and other environmental factors (e.g. soil NH4-N and NO3-N concentrations) was highly variable over the 5-yr study period. In the more deeply rooted shrub D. leiophyllum, annual precipitation was important in predicting ANPP, which was highest in the wettest year as a result of frequent large precipitation events. In the more shallow-rooted grass B. curtipendula, the magnitude of annual precipitation was not a key determinant of ANPP as frequent small precipitation events in the summer with relatively few long dry inter-pulse periods seemed to regulate periods of active growth. Supplemental winter water during a very dry winter also stimulated ANPP in B. curtipendula, suggesting the critical importance of winter precipitation in this grass, especially during a dry year. In the succulent O. phaeacantha, there was no relationship between ANPP and annual precipitation, but small precipitation events with short inter-pulse periods during the winter and fall may have generated the greatest productivity.
ANPP was regulated by soil NO3-N and NH4-N concentrations, particularly in wet years. In D. leiophyllum, soil NH4-N was positively correlated with ANPP in wet years but not in average or dry years. In B. curtipendula, soil NO3-N and NH4-N were positively correlated with ANPP in wet years compared with dry years. In O. phaecantha, precipitation magnitude and inter-pulse duration were positively correlated with ANPP. In average and wet years, c. 70–90% of the variability of ANPP could be explained by climatic and soil factors. In dry years, only 32% of the variability in ANPP could be explained by these same factors. Therefore, in dry years, other factors (e.g. herbivory, aboveground and belowground competition, and the pattern of plant recovery from drought stress) apparently have important impacts on plant productivity. Consequently, because of the diversity of environmental factors regulating ANPP in these three representative species, and their interactive effects, it may be difficult to accurately predict plant response in this desert ecosystem to variable timing and magnitude of precipitation, especially in dry years.