Figure Legends
Figure 1. Schematic representation of five mechanisms involved in the partition of the water inputs received by catchments towards transpiration and streamflow. The top panel shows three ways in which plant uptake and transpiration takes away water from streams. The first one (mechanism 1) involves the most acknowledged process of uptake from the unsaturated zone. The second one (mechanism 2) incorporates uptake from the saturated zone or the capillary fringe above it in focalized wetland and riparian areas with a closer contact between plant roots and water tables. The third one (mechanism 3) encompasses distributed uptake from the saturated zone or the capillary fringe in the matrix of the catchment. The second panel illustrates processes by which streams can take away water from plants reducing transpiration and enhancing stream flow. They include deepening streambeds (mechanism 4), which enhance hydraulic gradients and limit plant access to water tables in low areas or even in the matrix of the landscape; and burying riparian and wetland vegetation with sediment deposits (mechanism 5) which temporarily reduces transpiration in those highly transpiring zones.
Figure 2. The developing catchment of El Morro in Argentina. A. Map of the catchment illustrating the location of the major sub catchments as defined by their topographic limits, active permanent streams, and potential drainage lines. Groundwater monitoring sites, which in some cases include several wells, and meteorological stations are indicated. The dotted white frame shows the limits associated with the schematic view on the right. B. Scheme of new permanent streams indicating their period of formation (also referencing vanished or dead-ended segments). The area occupied by wetlands and sediment deposits is depicted together with the location of streamflow measurement points corresponding to the present study or to previously available records. The Quinto river is the pre-existing permanent water course receiving the new streams with a mean discharge of 5.5 m3 s-1.
Figure 3. Decadal groundwater level trends in the El Morro basin. The twenty seven sites are grouped according to their corresponding elevation belt within the catchment (higher > 700 m, intermediate 700-550 m, lower 550-500 m, terminal plain < 500 m). A zero level is assigned at the time of the first measurement in order to highlight the absolute level shifts starting from that point. Dotted lines illustrate periods and sites in which incisions were carved less than 500 m away or where sediment was deposited right on the location. Mean level changes per decade were calculated excluding the incision/deposit situations.  All sites correspond to water wells used for ranching whose static level was recorded in the past. In 11 of them, modern levels were measured directly in the same well or in a new ad-hoc borehole whereas in the rest levels (always < 1.5 m) were estimated by the presence of wetlands or lagoons.
Figure 4. Schematic transverse vertical cut of the Quebrachal and Río Nuevo sub catchments. The transect corresponds to the transition between the higher and intermediate belts (-33.464, -65.475 to -33.409 and -65.305, see Figure 2). The surface of the terrain corresponds to the central transect while belowground features were obtained using observation less than 600 m away from the transect. Depth of the crystalline basement comes from vertical electrical sounding (Barbeito 2008) and the old water table level from the long-term observation sites (Figure 3). The new water table level is reconstructed based on those same sites and landscape features like wetlands, lagoons and incisions. An aerial photograph together with google earth imagery were used to indicate the year of the advent of these landscape features.
Figure 5. Water table depth in 14 monitoring wells at the El Morro catchment. Levels were manually measured in hand augered wells. The locations of each pair or group of sites and their corresponding vegetation and relevant neighboring features are shown accompanying each plot. Dual numbering in the Y axes is color-coded like the markers in the plot and used to capture depth offsets between wells using an identical scaling across all plots.
Figure 6. Vegetation greenness, precipitation and water table depth dynamics for two periods of 84 days with contrasting hydrological conditions (recharge vs. discharge) at site A. Greenness is represented by an 8-day MODIS product, precipitation was recorded at the site and water table levels where measured with pressure transducers. Note the color-coded numbering in the Y axis matching the line color of each stand (green=wetland, black=5-day moving average in wetland, red=cropland). Grey zones represent the subperiod zoomed at the bottom.
Figure 7. Vegetation greenness, precipitation and water table depth dynamics for the whole study period in neighboring wells at site C (cropland and forest at position 2). Greenness is represented by an 8-day MODIS product, precipitation was recorded at “Coronel Alzogaray” meteorological station, 15 km away, water table levels were measured with pressure transducers. At the cropland an early gap on sensor data was filled with manual measurements (red dotted line). Note the color-coded numbering in the Y axis matching the line color of each stand (black=forest, red=cropland). Light grey bands depict each growing season while darker grey bands represent the subperiod zoomed at the bottom.
Figure 8. Base flow at different locations within the El Morro catchment. Grey areas depict isolated measurements before the continuing monitoring program started. Site location and baseflow synthesis are presented in Figure 2b and Table 1.
Figure 9. Greenness trends at sites with contrasting vegetation types/trajectories in El Morro catchment. NDVI from Landsat imagery for areas occupied since the beginning of the period by forests (n=11), croplands (n=16), and wetlands (n=20) where compared with areas experiencing the transition from croplands to wetlands (n=14) and fluvial deposits (n=7). A single case of wetland drying (site B) is also depicted. Alternate white and grey bars highlight successive growing seasons and letters indicate significant differences within years based on ANOVA. The lower panel shows mean monthly precipitation for the whole catchment and its 12-month moving average. The erosion episode took place in February 2015.