Multitaxonomic Diversity Patterns along a Desert Riparian–Upland Gradient

1. What they did

Riverside (riparian) areas are often said to hold more species than the drier land around them, but earlier studies disagreed, and most looked at only one group of animals or plants. The authors brought together surveys of 11 groups, all from the same stretch of the upper San Pedro River in southeastern Arizona, United States: forbs and grasses, shrubs, trees, solpugids, spiders, scarab beetles, butterflies, lizards, birds, rodents and mammalian carnivores.

The surveys had been started separately, for different purposes, between 1998 and 2005, so each group was sampled its own way: plot surveys and transects for plants, pitfall traps for invertebrates and lizards, quadrat counts for butterflies, point counts for birds, live-trapping for rodents, and transects for mammalian carnivores. Sampling covered four habitat types from the river outwards: river floodplain, river terrace, desert scrub near the river and desert scrub far from it.

For each group they compared:

• Species richness (how many species), adjusted for sampling effort.
• Abundance (how many individuals), standardised within each group.
• Community similarity (how many species two habitats share) and turnover between habitats.

They then combined the groups in a meta-analysis to see the pattern for the whole community, and tested whether five biological traits (trophic level, body size, life span, how the animal regulates body temperature, and whether it is a plant, invertebrate or vertebrate) explained why groups differed.

2. What they found

Group by group

• Groups differed. Comparing the four habitat types, trees, forbs and grasses, scarab beetles, birds and mammalian carnivores tended to be richest in the floodplain; shrubs, spiders, butterflies and rodents on the terrace; solpugids in the near-river desert scrub; and lizards equally on the terrace and near-river scrub. Only one of these differences was statistically significant: solpugids were richer in near-river scrub than in the floodplain.
• Trees and birds were significantly more abundant in the floodplain than in any other habitat type. Solpugids were significantly more abundant in both desert scrub types than in the floodplain.
• Comparing the two zones, scarab beetles, lizards and birds were significantly more abundant in the riparian zone, and solpugids in the uplands.

Shared species and turnover

• 24 of 56 habitat comparisons shared less than half their species. Similarity was highest between neighbouring habitat types (mean 0.61) and lowest between floodplain and far desert scrub (mean 0.31), though these differences were not statistically significant.
• Similarity between riparian and upland zones ranged from 0.166 for scarab beetles to 1.0 for mammalian carnivores, which used both zones.
• Turnover between habitats was greater than expected by chance for every group except mammalian carnivores, and significantly so for eight groups.

All groups together

• The riparian zone held significantly more species and more individuals than the uplands.
• The floodplain held significantly more individuals than the terrace, but not more species.
• The biological traits explained almost none of the differences between groups. The only significant result was for life span, in abundance between floodplain and terrace.

3. What they concluded

• How species richness and abundance change from river to upland differs between groups, even within one place. This may partly explain why earlier single-group studies disagreed.
• Taken together, the community is richer and more abundant in riparian zones, and riparian and upland zones hold quite different sets of species. This high turnover raises the total number of species in areas where riparian habitat is intact.
• The uplands also hold their own plants and animals. Conserving the whole gradient, from river to upland, matters for the area’s biodiversity.
• The traits they tested may not be the right ones, the number of groups (8 or 11) may have been too small, or traits may act in combination. Seasonal and year-to-year change in species also matters and should be studied alongside spatial patterns.

4. How tracking was or was not used

Mammalian carnivores were surveyed by tracks and signs along transects. The main paper says only “transects”; the details are in the online supporting methods. Ten routes, 1.9 to 11.8 km long, mostly loops along a road or trail in the scrub or grassland beside the river and back through the riverside forest, were each surveyed three or four times between October 1998 and January 2000, at least two months apart and at least two weeks after any rain. Tracks were identified by comparison with published field guides and with foot measurements from live-trapped and road-killed animals. Each sign was identified to species (to genus for hooded and striped skunks). To avoid counting the same individual twice on one transect, a sign was recorded only if it was more than 1 km from the last recorded sign of that species or in a different habitat type. All other groups were surveyed by trapping, plot searches or direct observation.

5. Tracker or observer certification

Not reported. Nothing is said about who identified the carnivore tracks and signs or how experienced they were. The bird surveys used nine experienced observers trained for at least 2.5 weeks each season, but that was for bird identification by sight and sound, not tracking.

6. Other technologies, skills or innovations

• Sign locations recorded with a Global Positioning System (GPS) receiver and a topographic map, and habitat type assigned with a Geographic Information System (GIS) vegetation layer.
• Meta-analysis used to combine very different survey methods into one community-wide result, rather than simply counting how many groups showed a pattern.
• Free diversity software: EstimateS, PARTITION, PopTools and MIX.

7. Implications for conservation (from the paper)

• Groundwater pumping for a growing human population is the main threat to the San Pedro River. How river drying would affect biodiversity depends on how much richer and how different the riparian zone is.
• The results underline the importance of conserving riparian areas and the water processes that sustain them.
• Upland zones have their own distinct plants and animals, so conservation is needed along the whole gradient. Without it, the high biodiversity of the San Pedro and similar watersheds would be much reduced.