Floodplain ecohydrology: Climatic, anthropogenic, and local physical controls on partitioning of water sources to riparian trees

Abstract

Seasonal and annual partitioning of water within river floodplains has important implications for ecohydrologic links between the water cycle and tree growth. Climatic and hydrologic shifts alter water distribution between floodplain storage reservoirs (e.g., vadose, phreatic), affecting water availability to tree roots. Water partitioning is also dependent on the physical conditions that control tree rooting depth (e.g., gravel layers that impede root growth), the sources of contributing water, the rate of water drainage, and water residence times within particular storage reservoirs. We employ instrumental climate records alongside oxygen isotopes within tree rings and regional source waters, as well as topographic data and soil depth measurements, to infer the water sources used over several decades by two co-occurring tree species within a riparian floodplain along the Rhône River in France. We find that water partitioning to riparian trees is influenced by annual (wet versus dry years) and seasonal (spring snowmelt versus spring rainfall) fluctuations in climate. This influence depends strongly on local (tree level) conditions including floodplain surface elevation and subsurface gravel layer elevation. The latter represents the upper limit of the phreatic zone and therefore controls access to shallow groundwater. The difference between them, the thickness of the vadose zone, controls total soil moisture retention capacity. These factors thus modulate the climatic influence on tree ring isotopes. Additionally, we identified growth signatures and tree ring isotope changes associated with recent restoration of minimum streamflows in the Rhône, which made new phreatic water sources available to some trees in otherwise dry years.

Key points: Water shifts due to climatic fluctuations between floodplain storage reservoirsAnthropogenic changes to hydrology directly impact water available to treesEcohydrologic approaches to integration of hydrology afford new possibilities.

Keywords: Rhône; climate change; oxygen isotopes (δ18O); soil moisture; tree rings; water partitioning.

Figure 1

Schematic showing how two species with contrasting rooting depths may record different annual values of δ¹⁸O based on variability in hydrologic partitioning between floodplain storage reservoirs. In the Rhône case, δ¹⁸O in Q is generally very similar to that in GW because hyporheic flow dominates the alluvial aquifer. The controlling parameter space is likely to vary with relative elevation in the floodplain, as well as the depth to gravel.

Figure 2

(a) Map of study area showing Pierre-Bénite within the great Rhône basin and France. Detailed maps of study site showing tree locations on an (b) aerial photograph and (c) LiDAR DEM backgrounds. The location of the piezometer used in Figure 6 is also shown, as is the River Kilometer used for the stage-discharge relationship in Figure 5. Source: BD Ortho©, BDT Rhône©, IGN.

Figure 3

Monthly precipitation δ¹⁸O averaged between Avignon and Thonon Les Bains (Figure 2) from the Global Network for Isotopes in Precipitation (GNIP—http://www-naweb.iaea.org/napc/ih/IHS_resources_gnip.html), demonstrating a seasonal cycle of increasing δ¹⁸O during the growing season. Note: δ¹⁸O precipitation value of −5.3‰ is computed as an average of growing season monthly averages (shown as dashed line).

Figure 4

Last 20 water years (September-October) of total precipitation and mean annual discharge data, including growing season (May-August) values. Vertical bars highlight distinct years in the record when we might expect differences in water availability and thus oxygen isotopes in tree rings.

Figure 5

(a) Stage-discharge relationship for River Kilometer (RK) 9.5 along the “Old Rhône” river channel (Figure 2). “High” and “low” trees elevation range is indicated in red and blue, respectively. The dashed line indicates the average floodplain surface elevation and the dotted line marks the average gravel layer elevation, as determined by penetrometer. (b) Daily discharge data for the study period (provided by the Compagnie Nationale du Rhône). Discharges required to exceed the gravel elevation for high and low trees are indicated (based on stage-discharge curve in Figure 5a).

Figure 6

Piezometer (P540, Figure 3) and river stage data (provided by the Compagnie Nationale du Rhône) illustrating offset between river level and floodplain water table elevation. Average gravel elevation, measured via penetrometer, is also depicted for high versus low trees. The elevated water table since 2000 is due to flow restoration.

Figure 7

Mean oxygen isotope ratios, δ¹⁸O, for high versus low trees for two co-occurring species: (a) Fraxinus excelsior and (b) Populus nigra. For Figures 10, the following apply. The year 2000 flow restoration is indicated, as well as individual years under closer scrutiny in this study, where the color coding corresponds to that in Figure 4. Stars near the x axis indicate significantly different values between the grand means of cohorts for individual years (based on two standard errors around the mean). N values listed in the legend are based on the number of trees used to compute cohort means.

Figure 8

Mean detrended, dimensionless ring width index (based on 30 year spline fit) for high versus low trees for two co-occurring species: (a) Fraxinus excelsior and (b) Populus nigra.

Figure 9

Mean oxygen isotope ratios, δ¹⁸O, for (a) Fraxinus and (b) Populus rooted deep versus shallow soils (measured by penetrometer to first refusal). (c and d) Mean δ¹⁸O for “high” versus “low” gravel elevations determined by subtracting penetration depths from floodplain surface elevation from determined from LiDAR.

Figure 10

Mean detrended, dimensionless ring width index (based on 30 year spline fit) for (a) Fraxinus and (b) Populus rooted in deep versus shallow soils (measured by penetrometer to first refusal). (c and d) Mean growth for each species rooted at “high” versus “low” gravel elevations determined by subtracting penetration depths from floodplain surface elevation from determined from LiDAR.

Figure 11

Individual oxygen isotope ratios and growth for each species in high spring snowmelt versus high spring rain years (see Figure 4). (a) Populus δ¹⁸O; (b) Fraxinus δ¹⁸O; (c) Populus growth; and (d) Fraxinus growth. Supporting statistics are provided in Table 1.

Figure 12

Individual oxygen isotope ratios and dimensionless growth for each species in wet versus dry years before and after the flow restoration. (a) Pre-restoration Populus δ¹⁸O; (b) post-restoration Populus δ¹⁸O; (c) pre-restoration Fraxinus δ¹⁸O; (d) post-restoration Fraxinus δ¹⁸O; (e) pre-restoration Populus growth; (f) post-restoration Populus growth; (g) pre-restoration Fraxinus growth; and (h) post-restoration Fraxinus growth. Supporting statistics are provided in Tables 1 and 2. Arrows indicate the approximate direction of change.