Decreased extraction despite crop intensification was largely enabled by increases in irrigation efficiency, including decreased water losses during transport to fields and basin-wide implementation of drip irrigation . Drip irrigation is known to produce much less off-field sediment transport during the irrigation season than sprinkler and furrow irrigation techniques. Thus, large scale conversion from furrow to drip irrigation led to decreases in irrigation water use, and may have also played a role in decreasing Salinas River suspended sediment concentrations in the latter 20th to early 21st centuries.Decreasing trends in suspended sediment concentration-discharge relationships were observed in the lower Salinas River from 1967-2011 despite increasing activities of wildfire and agriculture in the watershed over this period. Increases in effective burn area and total crop area have been generally found to increase sediment production at the watershed scale. Shifts in crop structure were dominated by a rise in row crops over this time period, which would also have been expected to increase sediment production. Row crop fields of are often left bare over the winter, rendering them prone to rainfall/runoff driven erosion, and degradation of the necessary drainage networks of earthen ditches can result in further increases in sediment export . With the exception of changes in irrigation practices, potential control of decreasing suspended sediment loads from other anthropogenic activities can be discounted due to limitedareal extent or timing. Urbanization increased,blackberries in containers but only to approximately 2% of the Salinas River watershed area.
While urbanization can lead to decreases in discharge-corrected CSS values by increasing the production of runoff from precipitation without concomitant increases in sediment production , no shift in the P–Q relationship was observed in the Salinas River between 1967 and 2011 . Conversely, the damming of Salinas River subbasins and attendant sediment trapping has been estimated to have significantly decreased sediment flux relative to pre-dammed conditions . However, dam emplacement in the Salinas Watershed occurred before the period of suspended sediment record, and the trapping characteristics of their reservoirs are not expected to have changed significantly over the intervening years . Wildfire activity was insufficient to counteract the negative inter-decadal trend in suspended sediment load, even though the years with the highest effective burn areas in the Salinas River watershed fell toward the end of the record. There was some indication that the large fires preceding the 1978 water year, coupled with the high Q-producing storm events of that year, may have increased suspended sediment load in agreement with the findings of Warrick et al. . However, other years with high effective burn areas and relatively high Q-intensities did not express consistent increases in CSS. The lack of wildfire control on inter-decadal scale trends in sediment loads may be due to issues of scale and the areal extent of burning. Often findings of dominant wildfire control come from studies of small, headwater catchments that have experienced burning over a high proportion of land area . Indeed, the suspended sediment flux from the Arroyo Seco subbasin was found by Warrick et al.to be highly controlled by the coincidence of wildfire and large storms on the basis of two instances of nearly complete burning of the watershed’s oak and chaparral shrublands.
The role of fire in maintaining chaparral vegetation communities and dominating sediment production in the semi-arid foothills and low mountains typical of coastal central and southern California has been extensively reported . In contrast, the undammed Salinas River watershed is an order of magnitude larger than the Arroyo Seco, with about half the average relief, and extensive agricultural development consisting of irrigation agriculture in lowlands and extensive grasslands on lower foothill slopes. As a result, even the largest EBA calculated for the Salinas River were only 10% of the undammed watershed . Thus the larger land area of the Salinas River watershed with numerous tributary drainages and divides and attendant mosaics of vegetation and microclimates has resulted in a mosaic of small wildfires relative to watershed size during any given fire season. Disconnected fire patches would be expected to produce less effective transfer of fire generated hillslope sediments to channels, in contrast to that of a completely burned watershed. Lowlands in the mainstem drainage network of the Salinas River also likely present a sink that moderates the signal of hillslope sediments produced from burned land surface, further obscuring the signature of more extensive burn years . Lavé and Burbank found a similar disappearance of wildfire control on inter-decadal scale sediment production when scaling up from small, 10-1 to 101 km2 scale headwater subbasins to 102 km2 scale watersheds in the San Gabriel Mountains of the southern California. Furthermore, semi-arid systems with intermittent flow and high discharge losses to groundwater recharge like the lower Salinas River tend to have longer residence times for suspended sediment due to increased incidence of in-channel deposition, particularly during flows into dry channels . Previous and continuing alterations to the Salinas River and its watershed may have exacerbated the attenuation of hillslope sediment production signals by destabilizing the channelized system of the lower Salinas and drawing down the ground water table.
Widespread deforestation along the banks of the lower Salinas River in the 19th century seemed to have decreased bank strength and led to a transition from a single meandering channel to a disorganized sandy active corridor, with localized and incipient braiding, which persists today. Intensive groundwater pumping remains above replacement , which could further exacerbates this scenario of lowland moderation of highland sediment production signatures by increasing the proportion of channelized flow abstracted to groundwater recharge . Indeed, early wet season flows have been observed to completely attenuate before reaching gauges S1 and S2, and thus completely depositing their suspended sediment loads into the channel . A more direct human cause of the overall negative trend in CSS, and the period of low CSSf that has persisted since the mid-1990s, is the conversion of agricultural operations to drip irrigation. Previously dominant methods of irrigation, particularly furrow, were known to produce large amounts sediment from off field transport and irrigation canal erosion . Drip irrigation has been shown to result in much lower off-field transport of sediment than sprinkler and furrow methods , and was introduced to California in the early 1960s. Large scale shifts in agricultural practices toward drip irrigation was contemporary with the decrease in suspended sediment concentration-discharge relationship observed for fine and sand sized sediment in the lower Salinas River. Although drip irrigation was used for only ~ 14% irrigated land coverage by 1993, over the next 17 years land area under drip irrigation quadrupled, replacing sprinkler and furrow methods as the primary irrigation practice in the Salinas Valley. This change in irrigation technology may be the dominant driver of not only the decreasing fine and sand sediment production trend found from 1967-2011, but also the timing of the departure from the hydrologic and climatic controls that Gray et al.found for fine sediment in the late 1990s to early 2000s. Drip irrigation has largely replaced older methods of irrigation for certain crops throughout California and other semi-arid or dry-summer climatic regions over recent decades, primarily due to increases in yields of high value row crops such as tomatoes . Widespread adoption of drip irrigation for such crops could have the unintended side-effect of reducing the entrainment of agricultural sediments into fluvial systems in these regions, which may have beneficial water quality consequences. Fluvial sediments are the greatest single impairment of rivers and streams in California,blackberry container many other parts of the U.S. and the world . Furthermore, agricultural sediments are often exposed to surface reactive nutrients such as phosphates, and multiple pesticides, many of which are hydrophobic and primarily transported off-site in association with fine sediments . Although winter season erosion remains an issue on such fields in single cropped areas, the time period between pesticide application and off-field is much longer for sediments eroded during the winter, perhaps decreasing winter off-field sediment associated pesticide fluxes. Thus, increases in drip irrigation use could yield a potential benefit in reducing the delivery of agricultural sediment to water bodies, particularly during the times when these sediments are most contaminated.It is well recognized that the inadequacy of conventional approaches for characterizing the key parameters and monitoring the key processes at over large enough areas yet with high enough resolution hinders our ability to optimally manage our natural water resources.
High resolution geophysical methods, such as GPR, hold promise for improved and minimally invasive characterization and monitoring of the subsurface. Here, we review several case studies where we have successfully used GPR for a variety of environmental and precision agricultural investigations. Section 2 focuses on the use of GPR for estimating parameters that are important for environmental and agricultural applications, such as hydraulic conductivity, sediment geochemistry, lithofacies zonation, and water content. Section 3 focuses on the use of time-lapse geophysical methods for assisting with remediation investigations, such as for to detecting biogeochemical hydrological processes that occur during remediation and the distribution of remediation amendments. This collection of case studies illustrates the utility of GPR for environmental and agricultural applications. Geophysical data are being increasingly used in hydrogeological site-characterization to obtain a better understanding of heterogeneity and its control on flow and transport. Such data can bridge the gap between the typically sparse conventional field characterization data and the need to realistically parameterize numerical transport models. In this section, we focus on the use of GPR to estimate hydrogeologic parameters that are important for agricultural and environmental studies, such as: water content, lithofacies zonation, hydraulic conductivity, and sediment geochemistry. Many of these studies also involved the development and use of stochastic estimation methodologies, which have enabled us to systematically fuse GPR and other sparse but direct measurements. Recent studies have shown that careful irrigation management can have beneficial effects on many crops, including almonds, citrus, prunes, pistachios and wine grapes. In particular, moderate water stress on grapevines early in the growing season can have a positive impact on grape quality. Thus, understanding when and how much irrigation to apply is critical for optimized wine grape production. Natural geologic processes, however, can cause soil variations and associated water-holding capacity to vary significantly, even over distances of a few meters. Given that the “industry standard” to vineyard soil characterization is to collect soil or water content measurements on a 75 m grid, grape growers typically do not have enough information about water content variations to guide precision irrigation. We have used GPR methods to estimate soil water content within agricultural sites in a non-invasive and manner and with high spatial resolution. Using 900 MHz GPR ground wave travel time data, we have estimated soil water content distribution in the top 15cm of the soil layers at high spatial resolutions and as a function of time at the Robert Mondavi Vineyard in California. Comparison with conventional ‘point’ soil moisture measurements, obtained using time domain reflectometry and gravimetric techniques revealed that the estimates of GPR-obtained volumetric water content estimates were accurate to within 1% by volume. The density of the obtained water content estimates was perhaps the highest density shallow moisture measurements obtained to date; the study produced 20,000 measurements of soil water content over the 3 acre study site. Water content distribution in deeper layers can also be obtained using GPR reflection arrivals if sufficient contrasts in dielectric properties exist. For example, Figure 1 illustrates the average volumetric water content estimated using data from 100 MHz GPR reflections associated with a subsurface channel, located 1.0-1.5 m below ground surface at a 2 acre Dehlinger Vineyard Chardonnay block in Sonoma County. Figure 1 shows how the channel has influenced subsurface water distribution, and suggests a correlation between water distribution and canopy density . At this site as well as the Mondavi site, it was clear that water content distribution was linked to soil textural and canopy vigor variations. Huisman et al. provide more information about the use of GPR for water content estimation. We have used cross-hole geophysical data to provide multi-dimensional estimates of hydraulic conductivity at a DOE bacterial transport site located near Oyster, VA.