Seasonal carryover of water and effects on carbon dynamics in a dryland ecosystem

Abstract

Net primary productivity in arid and semiarid regions is controlled by water availability for which rainfall has been a commonly used proxy at annual scales. However, the hydrological partitioning occurring through the water balance can also shape both seasonal and annual net ecosystem productivity. In this study, we used 10 years of water and carbon flux measurements in a mixed shrubland watershed of the Chihuahuan Desert to investigate the seasonal variability and controls on net ecosystem production. Over this period, the site exhibited a bimodal rainfall regime with an average of 211 and 67 mm for the wet (July–December) and dry (January–June) periods of the year, respectively. During the wet season, soil infiltration and channel transmission losses led to an average of 35.8 mm of water stored in the subsurface for subsequent dry periods. By contrast, dry seasons consumed 30.3 mm of stored subsurface water to fulfill the ecosystem water demand, particularly during the springtime. In response, gross primary productivity occurred in equal amounts for both seasons, while ecosystem respiration was substantially higher during the wet season. This resulted in the mixed shrubland acting as an annual net sink of carbon (average of 153.2 g C m−2 year−1) of which 65% occurred during the dry season. We determined that the wet season provided water in excess of vegetation demands in that season, a portion of which was stored in the watershed subsurface for subsequent use, leading to a legacy effect. Gross primary productivity during the dry season was dependent on carryover soil moisture accessed by deep‐rooted shrubs. Our coordinated observations of water–carbon dynamics revealed that water availability during the wet season can sustain the annual ecosystem productivity by impacting the subsequent dry season carbon uptake.

Non‐Stationary Lags and Legacies in Ecosystem Flux Response to Antecedent Rainfall

“…In arid ecosystems, evaporation is more likely to depend on antecedent rainfall to ensure there is moisture available. While evaporation is clearly inherently instantaneous (when water and energy are available), transpiration, particularly from deeper‐rooted species, could potentially provide mechanisms for memory to act (Pérez‐Ruiz et al., 2022). For instance, delays between rainfall and peak plant water content could buffer the response of transpiration to precipitation (Feldman et al., 2020).…”

Hillslope to channel hydrologic connectivity in a dryland ecosystem

“…This mechanism was identified based on: (1) infiltration depths after events that were limited to the upper 5–15 cm, (2) peak hillslope runoff amounts that were linked primarily to the maximum rainfall intensity ( I 30 ), and (3) the negligible effects of antecedent wetness (θ Sur ) on runoff production. The limited role of θ Sur was likely due to the high evapotranspiration amounts in the summer (~2–4 mm/day; Pérez‐Ruiz et al, 2022), leading to a rapid drying of the surface soils. Similar conclusions were obtained by Kampf et al (2018) for small watersheds in the Sonoran Desert.…”

“…The first‐order ephemeral channel drains a portion of the piedmont slope from east to west that emanates from the San Andres Mountains and is largely disconnected from deep water tables which are not subject to groundwater pumping (Schreiner‐McGraw & Vivoni, 2017). Local climate is classified as a cold desert (Koppen zone BWk), with an annual average precipitation of 278 mm and a mean annual temperature of 18°C (Pérez‐Ruiz et al, 2022), with most of the precipitation occurring during the North American monsoon (NAM; Adams & Comrie, 1997) between July and September. The ecosystem is a mixed shrubland consisting primarily of creosote bush ( Larrea tridentata ), honey mesquite ( Prosopis glandulosa Torr.…”

“…For the analysis of internal and outlet runoff flumes, we use the period of June 1, 2010, to September 15, 2021, after establishment of the environmental sensor network (Templeton et al, 2014). We divided each year in the record into two 6-month seasons: cool (October-March) and warm (April-September), based on monthly mean air temperature, with the latter containing storms during the NAM (Pérez-Ruiz et al, 2022).…”

Are Plant Functional Types Fit for Purpose?

“…This indicates a requirement for further ecosystem characteristics to model the connection between the land and the atmosphere, determining how input forcings to the land surface (i.e., climate) are transformed into outputs (i.e., carbon, water, and energy fluxes). Such information is likely to be related to vegetation, but may also describe site soil characteristics (De Long et al., 2019; Zhou et al., 2021), disturbance history (Amiro et al., 2010; Pugh et al., 2019), or factors such as hydrological functioning (Euskirchen et al., 2020; Griebel et al., 2020; Pérez‐Ruiz et al., 2022) and topography (Hoover et al., 2021; Xie et al., 2021). Modern ESMs nearly always account for such characteristics, through for example, soil type maps and hydrological modules, but analysis of results may not consider these sources of information as explanatory factors.…”