South Texas Sandy Savanna

This Tamaulipan ecological system is dominated by perennial grasses with sparse overstory of mesquite or oak trees and thornscrub. Stands of the system are typically dominated by Prosopis glandulosa in the overstory, which may be sparse, giving the aspect of an open grassland with scattered trees and shrubs. Or, more commonly, the system occurs as shrub-dominated patches within a grassy matrix. There will typically be an emergent canopy ranging to about 6 or more meters in height, composed of Prosopis glandulosa sometimes with Ebenopsis ebano and/or Celtis ehrenbergiana. Sometimes the overstory canopy is well-developed and would be considered woodland. These patches often coalesce to form significant expanses of shrubland. Dominant grasses are Cynodon spp. This system was once a common matrix system, but has largely been converted to desert scrub and exists as remnant patches. Degraded subtropical forests and woodlands may have similar structure but are not included in this system because different ecological processes maintain them.

This system is typically dominated by Prosopis glandulosa in the overstory, which may be sparse, giving the aspect of an open grassland with scattered trees and shrubs. Or, more commonly, the system occurs as shrub-dominated patches within a grassy matrix. There will typically be an emergent canopy ranging to about 6 or more meters in height, composed of Prosopis glandulosa sometimes with Ebenopsis ebano and/or Celtis ehrenbergiana. Sometimes the overstory canopy is well-developed and would be considered woodland. These patches often coalesce to form significant expanses of shrubland. Sites with somewhat tighter soils tend to have a denser shrub stratum, while deep sands and sandy sites tend to be more open, often with sizeable areas lacking significant shrub cover and dominated by a primarily graminoid herbaceous layer. The shrub component of woody patches or shrublands is commonly dominated by species such as Zanthoxylum fagara, Condalia hookeri, Celtis ehrenbergiana, Opuntia engelmannii var. lindheimeri, Diospyros texana, Colubrina texensis, Cylindropuntia leptocaulis, and Vachellia farnesiana (= Acacia farnesiana) (Elliott 2011). Prosopis glandulosa is almost always present, and is often dominant to codominant and occupies the highest canopy position (sometimes sharing that position with few other species), sometimes to 6 m in height. Numerous other species may also occur in the shrub layer, including but not limited to Schaefferia cuneifolia, Mahonia trifoliolata, Forestiera angustifolia, Lycium berlandieri, Aloysia gratissima, Salvia ballotiflora, and Ziziphus obtusifolia. The diversity of the shrub layer is significantly influenced by land-use history, with recently cleared areas sometimes being represented by a near monoculture of Prosopis glandulosa in the overstory, Pennisetum ciliare in the herbaceous layer, and Opuntia engelmannii var. lindheimeri as the most conspicuous component of the shrub layer. The herbaceous layer is typically dominated by graminoids and may be quite dense (60-100% cover). Grasses, such as Schizachyrium scoparium, Schizachyrium littorale, Chloris cucullata, Paspalum monostachyum, Paspalum plicatulum, Elionurus tripsacoides, Bouteloua rigidiseta, Urochloa ciliatissima, Heteropogon contortus, Eragrostis secundiflora, Bothriochloa laguroides ssp. torreyana, Trichloris pluriflora, Aristida spp., Sporobolus cryptandrus, and/or Dichanthelium spp., commonly dominate or codominate the herbaceous layer. Forbs are also common, including species such as Gaillardia pulchella, Eriogonum multiflorum, Croton spp., Cnidoscolus texanus, Aphanostephus skirrhobasis, Rudbeckia hirta, Verbesina encelioides, Clematis drummondii, Cynanchum barbigerum, Thymophylla pentachaeta, Justicia pilosella, Nama jamaicense, Monarda punctata, Palafoxia texana, Florestina tripteris, Zornia bracteata, Croptilon divaricatum, Rhynchosia americana, and Wedelia acapulcensis var. hispida (= Wedelia texana), though some of these species are restricted to the sandiest sites (Elliott 2011).

Examples of the system are found on thinner eolian sands on the western side of the South Texas Sand Sheet, as well as other sandy sites such as those of the Eocene sands of the Carrizo, Queen City, and Sparta formations. It may also be found associated with other formations, such as Oakville sandstone and other formations producing sandy residuum. Typical sites are level to gently rolling. This system occurs on sandy soils, including sandy, sandy loam, and loamy sands. Ecological Sites include sandy to sandy loam sites, such as those of the Sandy, Loamy Sand and Sandy Loam Ecological Sites (Elliott 2011).

Fire and drought are key ecological processes in this system. This system was modeled by Landfire (2007a) using three classes: early-, mid- and late-seral. The early-seral class (1-20 years) is dominated by perennial grasses. This class was maintained by frequent replacement fire (MFRI = 5 years) as the dominant disturbance type in this class. Droughts slow progression of this class to mid-seral class. This class is modeled to last 20 years; this duration is extended due to limited mesquite seed dispersal mechanisms historically (prior to livestock introduction) (Landfire 2007a).

Mid-seral class (21-50 years) is the early development of shrub patches, often surrounding a mesquite trees. Tree canopy is sparse, but shrub cover is dense. Herbaceous cover is declining due to increased shrub and overstory canopy. Replacement fire is modeled to occur with a 20-year return interval. A mixed fire is modeled to occur with a 7-year return interval. Twenty-year drought is modeled to slow successional progression to late-seral class. The mechanism for drought effect may be an enhanced effect of fire. This class is modeled to last 30 years (Landfire 2007a).

The late-seral class (51+ years) is a closed-canopy, late-development stage that represents the continued development of shrub patches as they coalesce into more well-developed woodlands of Prosopis glandulosa (Archer 1989). In these late stages other species begin to colonize into woodlands and shrublands. Species present in mid-seral class are still present in late-seral class, but other species begin to colonize, such as Mahonia trifoliolata, Schaefferia cuneifolia, and Lycium berlandieri. Replacement fire is modeled to occur with a 200-year return interval. A mixed fire is modeled to occur with a 20-year return interval. Twenty-year drought is modeled and may slow increase in patch size but does not cause transition (Landfire 2007a).

The natural range of variation in disturbance within this vegetation is difficult to assess currently, because of dramatic changes resulting from severe overgrazing and the resultant changes in vegetation dynamics in the region which occurred in the early to mid-1800s. While most experts agree that this was a major habitat type of the region, the historic extent of mesquite savanna is arguable. Periodic fire, probably resulting from human sources of ignition, likely maintained the habitats as an open savanna. The average fire-return interval is 6 years. Periods of overgrazing apparently led to an alternative stable state in which fire does not play a significant role, and the habitat has become a closed shrubland community with little to no opportunity for reverting to mesquite savanna (Landfire 2007a). Many sites are currently occupied by denser shrub cover than historical condition (Landfire 2007a).

Threats from development, including development for agriculture and overgrazing by livestock, continue to convert or degrade existing stands. Road building and power transmission lines continue to fragment vegetation and provide vectors for invasive species. Persistent drought may result in loss of key species. Conversion of this type has commonly come from agricultural practices. Common stressors and threats include fragmentation from roads, agriculture and development, and non-native species invasion. Other stressors and threats include overgrazing/browsing by livestock, and possibly loss of pollinators.

According to Climate Wizard (TNC 2013), in 2050 global climate change model (using Medium A1B emission scenario and Ensemble Average general circulation model), the average annual temperature is predicted to rise approximately 5°F and average annual precipitation will not significantly change (TNC 2013). Seasonal shifts in precipitation predict increased fall (monsoon) moisture with similar levels of precipitation to current in the rest of the year (TNC 2013). Potential climate change effects on vegetation could include a shift to species adapted to a hotter, generally drier environment. While average precipitation amounts may remain similar or slightly decrease during the winter, spring and summer months, that, along with increased temperatures, may cause vegetation to experience less effective precipitation and more soil moisture deficit during much of the growing season reducing plant growth and increasing mortality from extreme events including exceptional drought. If the increased fall precipitation is from intense storms such as hurricanes, we can expect more disturbances from flooding and water erosion.

Examples of the system are found on thinner eolian sands on the western side of the South Texas Sand Sheet in Texas and related areas of Mexico.