Impact of adjacent land use on the ecological condition of riparian habitats: The relation between condition and vegetation properties

METHODS

Sampling sites

Sampling sites were established along the riverbank by considering three different adjacent land-use practices, including the following: forest (FOR, N=3), agricultural (AGR; N=5 sites) and urban areas (URB; N=4). Riparian vegetation in the FOR land-use type was in contact with old-growth remnants of tropical deciduous forest or secondary vegetation present for a length of at least 500 m perpendicular to the riparian vegetation. Crops adjacent to the AGR sampling sites were maize (Zea mays), sorghum (Sorghum bicolor), grass (Cenchrus ciliaris) and/or strawberry (Fragaria ananassa). Sites in the URB were located within semi-urbanized rural as well as urban areas. The different vegetation sampling sites were established far from rapidly flowing sections and/or meanders in the river to avoid variation in the vegetation structure caused by these ecological factors.

Riparian condition index

The ecological condition of each sampling site along the riverbank was determined by using a previously tested riparian condition index (RCI), with modifications according to our study site conditions (Jansen & Robertson, 2001). Each sampling unit consisted of a 500-m transect established along the riverbank and randomly located on one side of the river. The RCI considered six different biophysical, vegetative and landscape traits (subindex): (1) habitat continuity and width (HABITAT), (2) vegetation cover and structural complexity (COVER), (3) bank stability (BANKS), (4) standing and fallen debris (DEBRIS), (5) dominance of native vegetation versus exotic (NATIVES), and (6) the natural regeneration of woody seedlings (REGENERATION). As shown in Table 1, 17 indicators of the riparian condition were measured in the field to quantify the contribution of each subindex score to the overall condition index value.

Each indicator was weighted according to its ecological importance in each subindex, as given in Table 1. A lower value was assigned to the NATIVE and REGENERATION subindices. In the case of NATIVES, there is little information regarding how exotic plant species may perform ecological functions compared with the native plant species they have replaced (Jansen & Robertson, 2001). Therefore, the relative contribution of NATIVES in our index of ecosystem condition was lower than other subindices (Table 1 in the source publication). A low value was also given to the REGENERATION subindex since the presence of common species is highly variable across space and time and depends on a complex interaction of hydrologic and geomorphic processes that shape seedling establishment (González et al., 2018). Hence, this subindex may not be as relevant for the ecological condition assessment of riparian forests as other ecological indicators. For example, the incidence of rocks is considered an important component for the maintenance and stability of the riverbanks and for erosion reduction, so although Jansen and Robertson (2001) did not use this indicator in their index, the presence of rocks was included here as a useful indicator for assessing bank stability (e.g., Heartsill-Scalley & Aide, 2003).

The evaluations for each indicator were averaged for each site, scored, and weighted, then added to obtain a final score per site. Potential scores were from 0 (worst condition) to 50 (best condition). To summarize the results, the scores of the RCI were grouped into five categories: very poor condition (<25); poor condition (>25-<30); regular condition (> 30 - < 35); good condition (> 35 - < 40); and excellent condition (> 40).

Evaluations methods

The same observer performed all the evaluations regarding the ecological conditions to reduce any bias. Observations took place before the rainy season. Thirteen of the 17 riparian indicators were measured systematically in four 20 × 5-m perpendicular riverbank transects. These transects were located 125 m apart along the sampling unit to measure: (i) the canopy cover (%), which consisted of all trees > 5 m height; (ii) the understory cover, which included herbs, grasses, shrubs and juvenile trees from 1 to 5 m in height; (iii) the ground cover or low understory stratum, including herb and grasses < 1 m tall; (iv) the number of layers of vegetation; (v) the stability of the riverbank; (vi) the presence of boulders and stones; (vii) the incidence of standing dead trees; (viii) the density of terrestrial woody debris (> 10 cm diameter); (ix) the percentage of native species in the canopy, (x) the percentage of native species in the understory; (xi) the percentage of native species in the ground stratum; (xii) canopy regeneration measured as the density of tree seedlings ≥30 cm and ≤ 100 cm in height in the understory; and (xiii) understory regeneration referred to the density of shrubs seedlings ≥30 cm and ≤ 100 cm in height.

Furthermore, (xiv) the width of the riparian vegetation (on the side of the river being assessed) was measured at 10 evenly spaced points within each sampling unit; (xv) the leaf litter cover on the ground was estimated at 10 perpendicular transects (20 × 5 m) to the riverbank. Finally, (xvi) the longitudinal continuity of the riparian vegetation, and (xvii) the aquatic woody debris were assessed once through the entire 500-m sampling unit. For longitudinal continuity, a diagram was drafted where discontinuities in vegetation cover were identified along the 500-m sampling unit. For aquatic woody debris, four categories of tree and branch density along the 500-m bank section were made (Table 1 in the source publication).

Assessment of vegetation properties

At each of the 12, 0.1-ha sites, we established 10, 20 × 5 m (100 m2) transects on one side of the river, which were randomly positioned. At each sampling site, we recorded the density of stems and individuals of trees and shrubs with trunks ≥ 2.5 cm in diameter at breast height (DBH at 1.30 m); basal area (m²); mean height, from the maximum height recorded for all the individuals and the proportion of multi-stemmed individuals. Species were identified to the lowest taxonomic level in the field, and botanical samples were taken to an herbarium for those unidentified species. We recorded the total number of species (S) and the number of species that occur in a only one sample (unique number of species; U) (Magurran & McGill, 2011). One nonparametric richness index (the incidence-based cover estimator, ICE) and alpha diversity indices were also obtained. The Shannon index (H´), calculates the diversity of a site considering the proportional abundance of its species. The Simpson index (D), is an index of dominance where the diversity of the sample declines as the value of D increases. All indices were calculated with the EstimateS 9.1.0 program (Colwell & Elsensohn, 2014) by using in each case 200 permutations without replacement. The level of patchiness was set to 0 in order to prevent bias in estimating species richness because of clustering of the species themselves (Colwell & Coddington, 1994).