An examination of the potential wetland development landscape around managed reservoirs in the central U.S. Great Plains

Applied Geography 93 (2018) 16–24

Contents lists available at ScienceDirect

Applied Geography journal homepage: www.elsevier.com/locate/apgeog

An examination of the potential wetland development landscape around managed reservoirs in the central U.S. Great Plains

T

Kaitlyn M. Loefflera,c, Vahid Rahmanib,c,∗, Jude H. Kastensc, Donald H. Hugginsc a

Department of Civil, Environmental, and Architectural Engineering, University of Kansas, Lawrence, KS 66045, USA Department of Biological and Agricultural Engineering, Kansas State University, Manhattan, KS 66506, USA c Kansas Biological Survey, University of Kansas, Lawrence, KS 66047, USA b

A R T I C L E I N F O

A B S T R A C T

Keywords: Bathymetry Lake level LiDAR Reservoir Slope Wetland

Reservoirs around the world are losing storage capacity due to sediment infilling, leading to reductions in the quality and value of reservoir uses. However, the sediment accumulating in the upper ends of reservoirs, particularly around primary inflows within well-defined floodplains, could potentially be developing into wetland ecosystems that provide services such as sediment filtration, nutrient sequestration, and wildlife habitat. In this study, we examine lake level, wetland, and topographic characteristics for 20 large, federally operated reservoirs in the state of Kansas located in the central U.S. Great Plains. First, daily water level data and empty-basin topography (created by merging LiDAR elevation with bathymetry) are used to delineate the primary fluctuation zone for each reservoir, which we define to be the range of extents between the 25th and 75th lake level percentiles. Next, we use data from the U.S. Fish and Wildlife Service National Wetlands Inventory (NWI) to characterize wetland composition in the upper fluctuation zone (extent range between the 50th and 75th lake level percentiles) and the upper flood pool (extent range between the 75th percentile and the maximum designed flood pool elevation). For the final assessment, we examine the relationship between ground slope and NWI in the upper fluctuation zone and the upper flood pool. Results indicate that relatively low-sloped ground is classified as wetland in the upper fluctuation zone at 24% greater frequency than in the upper flood pool.

1. Introduction While man-made reservoirs and natural lake ecosystems are similar in their water usage, they differ in characteristics such as drainage area and age. Specifically, reservoirs are considerably younger than natural lakes and usually have larger drainage areas due to choice of location and purpose such as flood control (Cooke, Welch, Peterson, & Nichols, 2016; Hayes, Deemer, Corman, Razavi, & Strock, 2017). Soballe and Kimmel (1987) found that the ecological structure and function of rivers, river impoundments, and natural lakes on a broad scale varied along a composite gradient that changed with water residence time, drainage area, water depth, flow, and water clarity. Lakes and rivers occupied opposite ends of this spectrum with reservoirs typically occupying an intermediate position. Ecological structure maturity differences between natural lakes and reservoirs may reflect the artificial changes in water level and the disequilibrium state of reservoirs, given their anthropogenic origin, young age, and in some cases, regimented management. It is generally understood that in natural lakes the spatial variation of many physio-chemical and biological factors are related to shoreline ∗

length, depth, and wind-driven currents (Thornton et al., 1981). By contrast, these attributes appear to be of less importance in reservoirs, where the prominent determinants of observed spatial gradients in physio-chemical and biological conditions are relatively large riverine inflows and depth gradients characteristic of damming the river channel to create an in-line impoundment (Lehner et al., 2011). These upstream-downstream gradients in depth and flow often result in measurable gradients in turbidity, mixing depths, nutrient concentrations, primary production, and fish standing stocks along with other characteristics (Kennedy, Gunkel, & Thornton, 1982; Kimmel & Groeger, 1984; Kimmel, Lind, & Paulson, 1990; Lind, 1984). The uppermost riverine zones of the reservoirs are characterized as shallow, light-limited and high nutrient zones, whereas the deeper, clearer water area near the dam functions more similarly to natural lakes (Kennedy, Thornton, & Ford, 1985). As impoundments age, a common problem impacting reservoir management and sustainability is sediment infilling (deNoyelles and Kastens 2016; Rahmani et al., 2018; Schleiss, Franca, Juez, & De Cesare, 2016; Stene, 1946). This is particularly problematic in regions where precipitation, soil properties, topography, and land use all contribute to

Corresponding author. Department of Biological and Agricultural Engineering, Kansas State University, 920 N. 17 Street, 1016 Seaton Hall, Manhattan, KS 66506, USA. E-mail address: [email protected] (V. Rahmani).

https://doi.org/10.1016/j.apgeog.2018.02.010 Received 12 August 2017; Received in revised form 31 January 2018; Accepted 13 February 2018 0143-6228/ © 2018 Elsevier Ltd. All rights reserved.

Applied Geography 93 (2018) 16–24

K.M. Loeffler et al.

persist. This project aims to examine potential wetland development areas along the perimeters of federal reservoirs in Kansas to facilitate conservation and management of these potential resources for their ecological goods and services to society. The United States Army Corp of Engineers (USACE) offers legal protection to natural wetlands under Section 404 of the Clean Water Act if the wetland has hydrological characteristics, hydrophytes, and hydric soils (Johnson, 1992). The objectives of this study are to 1) use water level management data and topography to delineate the primary zone of potential wetland formation around the reservoir perimeter, 2) use data from the U.S. Fish and Wildlife Service National Wetlands Inventory (NWI) (USFWS, 2016) to characterize wetland composition in the reservoir flood pool, and 3) examine the relationship between ground slope and NWI occurrence in the reservoir flood pool. We use the NWI because it is currently the only documented source of delineated wetlands in Kansas and is a widely used and well-known standard despite uncertainties regarding its accuracy and completeness.

high levels of soil erosion (García-Ruiz et al., 2015; Ziadat & Taimeh, 2013). One type of soil erosion specifically expected to be affected by climate change is gully erosion, in which increased and more frequent runoff events create favorable conditions for gully development (Nearing et al., 2004; Poesen, Nachtergaele, Verstraeten, & Valentin, 2003). Gully erosion from agricultural lands has been acknowledged as a major supplier of the total sediment loads flowing into reservoirs, indicating the importance of land use on sedimentation rates (Fox et al., 2016). The disproportionately large amount of sediment brought in from cropland is especially important where farming communities make up a substantial portion of a reservoir's watershed. One such area is the state of Kansas located in the central United States, where cropland (approximately 50%) and grassland (approximately 42%) dominate the landscape (Peterson, Egbert, Price, & Martinko, 2004). With this land cover composition, and with its many large impoundments, Kansas is a good candidate for studying the impact of sedimentation on potential wetland formation around the upper ends of reservoirs (Rahmani et al., 2018). In addition to the inevitable aging and infilling process (deNoyelles and Kastens 2016), sediment infill rates may be increasing due to changes in the occurrence of extreme storm events (Coppus & Imeson, 2002; Mouri, Shiiba, Hori, & Oki, 2011; Valentin, Poesen, & Li, 2005). Kansas has seen an upward trend in the magnitude and frequency of extreme rainfall events during 1890–2013, with an even greater upward trend observed for 1981–2010 (Rahmani, Hutchinson, Harrington, & Hutchinson, 2016). With reduced water storage capacity and high sedimentation rates in the upper ends of many Kansas reservoirs, other impacts such as habitat loss (or gain) and water quality changes are being examined (Hargrove, 2008; Juracek, 2015). One impact of sediment infilling is the increased area of shallow water and total amount of nutrient-rich sediments that go along with it (Cooke, Welch, & Peterson, 2013). It is well documented that the quality of water found in our water bodies has decreased due to the increased nutrients in agricultural runoff (Verhoeven, Arheimer, Yin, & Hefting, 2006). Shallow water and increased nutrient loads are two parameters often found in wetland ecosystems compared to natural water bodies. With land use change and climate change, particularly in areas also experiencing an increase in larger storms, reservoir hydrological and ecological characteristics may change more rapidly (Raje & Mujumdar, 2010; Singh, Sinha, & Sankarasubramanian, 2014; Soundharajan, Adeloye, & Remesan, 2016). To help maintain the storage capacity of the reservoirs, the most immediate management plan is dredging, which is significantly costly. For example, $20 million (U.S.) was expended to remove 2.3 million cubic meters of sediment from John Redmond Reservoir in Kansas in 2016, which added approximately three more years of life for the reservoir (KWO, 2015; Rahmani et al., 2018). The cost of disposing dredged sediment can also be high, especially if the sediment contains harmful chemicals or trace metals such as arsenic, copper, lead, or mercury (Hargrove, Johnson, Snethen, & Middendorf, 2010). A possible alternative management strategy for these reservoirs whose upper end riverine zones are becoming shallower could be to manage these areas as wetlands if wetland characteristics exist. By filtering inflow through wetlands, in addition to improving water quality, this approach could slow the transmission of sediment into the main body of the reservoir where the storage capacity typically is most needed. According to the map coverage in Schenck, Wedel, and Monda (1992), most wetlands on cropland occur in the eastern one-third of Kansas. Additionally, most wetland types listed for Kansas primarily occur in far eastern Kansas, and the south-central and west-central portions of the state. It has been estimated that almost half of the historic wetlands in Kansas have been lost through drainage and agricultural land conversion (KGS, 2017). While it is unclear exactly where all of these wetlands were lost to agriculture, it is reasonable to assume that many were located in riparian and floodplain areas of the east where much of the land is farmed and many remaining wetlands

2. Methods and materials 2.1. Study area The study sites consist of the 20 federally operated reservoirs in the state of Kansas that had both LiDAR elevation data (to represent topography outside the reservoir) and bathymetry data (to represent lakebottom topography inside the reservoir) at the time of the analysis (Fig. 1; there are 24 federal reservoirs in Kansas, leaving four that did not satisfy these criteria). Reservoir watersheds for these 20 reservoirs consist mostly of grassland (58% of watershed area) and cropland (37%) (Martinko et al., 2014). General physical characteristics of this set of reservoirs are provided in Table 1. Two distinct climates cover Kansas, ranging gradually from a humid climate in the east to a semi-arid climate in the west. Average precipitation rates vary from approximately 1150 mm in the southeastern portion of the state to approximately 500 mm in the far west. The average temperature patterns go from the warmest in the southeast to the coldest in the northwest, with a statewide average low of 0 °C in January and an average high of 27 °C in July. All of the study reservoirs are located in the central and eastern part of the state, where precipitation ranges from approximately 600–1000 mm annually; however, the watersheds for the central reservoirs extend into the western portion of Kansas where there is less precipitation annually (Goodin, Mitchell, Knapp, & Bivens, 1995; Rahmani, Hutchinson, Harrington, Hutchinson, & Anandhi, 2015). 2.2. Hydrological analysis Historic water level elevation data and reservoir basin topography (a blend of LiDAR digital elevation data and bathymetry information) were used to determine the zone around each reservoir that captures the range of extents between typical dry conditions (25th percentile water level) and typical wet conditions (75th percentile water level), which we term the fluctuation zone. Within the upper fluctuation zone (the region between the 50th and 75th percentiles) and the upper flood pool (the region between the 75th percentile and the designed maximum flood pool level) we then examined median ground slope values summarized across wetland delineations from the NWI dataset. The historical reservoir water level data provided the range of typical water level elevations during wet and dry periods. These data were requested or retrieved from the USACE (USACE 2016a; USACE 2016b), the U.S. Geological Survey (USGS) (USGS 2016), and the U.S. Bureau of Reclamation (USBR 2016). For nine of the reservoirs, data were used starting the day that the reservoir first reached its regulation level through 2015. For ten of the reservoirs, data were used from the beginning of digital record keeping in 1995 through 2014 or 2015. For 17

Applied Geography 93 (2018) 16–24

K.M. Loeffler et al.

Fig. 1. Locations for the 20 federally operated reservoirs in Kansas that comprise the study sample. Imagery source: ESRI.

Table 1 Summary of characteristics for the 20 reservoirs examined in this study.

Table 2 25th, 50th, 75th, and 100th percentile water level elevations (m) along with the elevation of the top of the flood pool for the 20 study reservoirs.

Parameter

Max

Min

Mean

Median

Surface areaa (km2) Volumea (km3) Age (years, as of 2017) Maximum deptha (m) Shoreline lengtha (km) 25th percentile elevation (m) 50th percentile elevation 75th percentile elevation 100th percentile elevation Top of flood pool elevation

63.0 0.46 69 23.8 161.1 570.5 575.6 577.1 578.3 586.5

4.5 0.02 36 3.6 28.8 242.0 242.4 242.9 253.0 251.5

25.0 0.14 51 13.2 75.3 356.5 357.3 357.7 364.5 365.4

23.2 0.10 52 12.8 69.5 348.5 348.9 349.3 336.4 336.0

a

At normal operating water levels.

John Redmond, data were used from only 2013 and 2014 due to a recent change in regulation pool level (USACE 2013). The reservoir regulation level, also sometimes called the conservation pool level, is defined as the maximum water level elevation during normal operating conditions, without considering flood control storage (NOAA 2017; TWDB 2017). The daily water elevation levels were ordered from lowest to highest, and from this list we calculated the 25th, 50th, and 75th percentile water level elevations over the respective periods of record. These calculations provided reasonable estimates for water level during typical dry, normal, and wet periods, respectively (Table 2). The 25th percentile lake level corresponds with the maximum area that is inundated at least 75% of the time (typical low-water condition), while the 75th percentile lake level corresponds with maximum area that is inundated at least 25% of the time (typical high-water condition). The 50th percentile water level elevation was compared to the regulation level for each reservoir to determine if the median observed lake elevation was near the regulation level. Median lake levels for 17 of the study reservoirs were found to be within 0.3 m of their regulation

Reservoir

25th percentile

50th percentile

75th percentile

100th percentile

Top of flood pool

Big Hill Cheney Clinton Council Grove El Dorado Elk City Fall River Hillsdale John Redmond Kanopolis Kirwin Marion Melvern Milford Perry Pomona Toronto Tuttle Creek Webster Wilson

261.2 433.0 266.7 387.7 407.7 242.0 289.2 279.0 317.5 446.1 522.9 411.3 315.2 348.5 271.5 296.6 274.8 327.3 570.5 461.5

261.5 433.4 267.0 388.3 408.1 242.4 289.3 279.4 317.8 446.4 526.3 411.6 315.7 348.9 271.9 297.0 275.0 327.9 575.6 462.0

261.6 433.5 267.4 388.4 408.3 242.9 289.8 279.7 318.2 447.0 527.3 411.7 316.1 349.3 272.4 297.5 275.4 328.8 577.1 462.3

263.9 435.8 272.1 393.3 410.7 253.0 300.8 283.1 325.9 459.0 529.6 413.4 321.2 360.3 280.8 304.4 284.5 346.8 578.3 472.0

264.5 435.7 275.4 393.0 410.8 251.5 301.1 283.8 325.6 459.8 535.8 414.2 322.3 358.6 280.7 305.8 283.8 346.3 586.5 473.8

levels, while the median levels for three reservoirs located in the central part of the state (Kanopolis, Kirwin, and Webster) were considerably lower than their regulation level (by 0.45–4.27 m). Water levels for all of these federal reservoirs are highly managed; however, the reservoirs in the central part of Kansas are less likely to reach their regulation level due to the lack of rainfall and insufficient inflow.

18

Applied Geography 93 (2018) 16–24

K.M. Loeffler et al.

Fig. 2. A) 2-m LiDAR bare earth elevation raster for Clinton Lake, B) bathymetry (rasterized and up-sampled to 2 m) for Clinton Lake, and C) merged LiDAR and bathymetry to create the empty basin topography for the flood pool.

the original LiDAR topographic datasets (i.e. not blended with bathymetry, which is much less precise). These areas were further reduced by removing pixel footprints with zero slope, which almost invariably correspond with hydro-flattened locations and thus have already been identified as waterbodies during the outside production of the LiDAR elevation raster (Kansas LiDAR standards currently require hydro-flattening of all waterbodies at least 0.3 ha in size at the time of data collection, which was applied across most of the study area). NWI features were clipped to these same modified flood zones. Next, for each reservoir individually and also for all reservoirs combined, the median slope across pixels underlying NWI features was calculated for both the upper fluctuation zone and the upper flood pool (Fig. 3). The median slope from the upper flood pool NWI was subsequently applied as a threshold to determine the NWI area fraction within the upper fluctuation zone with lesser or equal slope.

2.3. LiDAR and bathymetry analysis LiDAR raster elevation data with 2-m resolution were obtained for the study area from the State of Kansas GIS Data Access and Support Center (DASC, 2016). These data are at least Quality Level 3 and have general vertical accuracy of ± 18.5 cm with 95% confidence (Heidemann, 2012). Gridded bathymetry data were supplied from studies conducted by the Kansas Biological Survey, USACE, USGS, and the Oklahoma Water Resources Board for the 20 study reservoirs (bathymetric reports for most can be found at https://kwo.ks.gov/ reservoirs). Reservoir footprints for the study sites were identified in the LiDAR data as hydroflattened extents (which are artificially leveled water surfaces, a common LiDAR post-processing enhancement) and replaced with bathymetrically derived lake bottom elevation so that the elevation dataset represented empty basin conditions. Fig. 2 shows an example of the process for Clinton Lake. For each reservoir, a flood pool boundary was developed based on the design specifications for the maximum supported water level (the primary purpose for all of the study reservoirs is flood control). Note that the flood pool level for these reservoirs is a fixed elevation that corresponds to the intended maximum volume of flood water that can be held within these reservoirs; for some reservoirs, this has actually been exceeded by the 100th percentile water level (Table 2). All further analysis was restricted to the interior of the flood pool polygons. The hydrological data were then used to create additional boundaries within this basin corresponding with the 25th, 50th, and 75th percentile water level elevations.

3. Results and discussion 3.1. Water level fluctuations The regions contained in between the 25th and 50th and between the 50th and 75th percentile boundary layers show the coverage areas under typical low-water and high-water conditions, respectively. These boundaries were used to calculate the regime-specific coverage area for each reservoir. Fig. 4 illustrates the coverage areas for Elk City and Webster reservoirs as two examples of the 25th, 50th, and 75th percentile boundary layers. Table 3 shows the various coverage areas for each reservoir, along with the percentage of area gained or lost relative to the 50th percentile area. The smallest absolute difference between the 25th and 50th percentile lake extents was 18 ha in Fall River Reservoir, while the largest difference was 621 ha in Kirwin Reservoir. Likewise, the smallest difference between the 50th and 75th percentile lake extents was 4 ha in Big Hill Reservoir, while the largest difference was 727 ha in Tuttle Creek Reservoir. The areal and vertical differences between the 25th and 50th and the 50th and 75th percentile water level elevations can be partially explained by the inflows to the reservoirs and the ground topography. For example, Fall River Reservoir is in the southeastern portion of the state where the highest amount of precipitation occurs. Since the reservoir receives relatively high inflow, it is frequently at its regulation level, causing the 25th percentile water level elevation to be near the

2.4. Potential wetland development areas Since the LiDAR data are very accurate and precise while the NWI data are of questionable precision, accuracy, and completeness, even modest assessments must be exercised with caution. A basic question that can be asked is if relatively low-slope ground in the upper fluctuation zone has a greater propensity to be classified as NWI wetlands than in the upper flood pool, which likely better resembles the more natural hydrology of the unimpounded portions of the river valley. This could support the notion that the upper fluctuation zone is more amenable to wetland development than areas outside the upper fluctuation zone. Using ESRI ArcMap and MATLAB software, percent-slope values were calculated for the flood zone (50th percentile to flood pool) using 19

Applied Geography 93 (2018) 16–24

K.M. Loeffler et al.

Fig. 3. NWI wetlands around Clinton Lake. Background is 2015 USDA-FSA National Agriculture Imagery Program (NAIP) data.

three main features that distinguish wetlands from other types of ecosystems. For example, different types of vegetation are able to thrive, or at least grow, in different zones of flooding (Keddy & Fraser, 2000). Constant flooding in a wetland is not ideal because emergent plants cannot grow and only submergent plants will be present. The opposite is also true: not enough water and only upland vegetation will be present. Therefore, a mix between constant flooding and no flooding will yield the most biodiversity in a wetland. Yearly water level fluctuations create vegetative biodiversity by killing off low-water shrubs during times of high water and allowing dormant species to regenerate from the wetland's seed bank (Keddy & Fraser, 2000). In addition, the extent, frequency and duration of water level fluctuations influence the physical processes of a lake, specifically erosion and sedimentation (Leira & Cantonati, 2008). Since water level fluctuations have such a profound impact on lake ecosystems, it is important to understand how human impacts, such as reservoir water level regulation, affect the reservoir ecosystem as a whole (Coops & Havens, 2005; Coops, Beklioglu, & Crisman, 2003).

50th percentile water level elevation. By contrast, Kirwin Reservoir has its watershed in drier western Kansas and thus receives relatively little inflow, and also has high evaporation. Therefore, large fluctuations of water level elevations during typical low-water conditions cause a substantial difference between the 25th and 50th percentile water level elevations. Generally, if water is plentiful, maintaining regulation level is easier and high-water fluctuations are more varied. If water is scarce, maintaining regulation level is more difficult and low-water fluctuations are more varied. Other factors such as topography help explain deviations from these precepts. As an example of the influence of ground topography, Cheney Reservoir is situated in a steeper-sloped valley than most of the other reservoirs, which makes for a more compact range of lake extents. Across the study sites, the ratio of area gained or lost relative to the 50th percentile indicates that from 2.0% (Fall River Reservoir) to 45.1% (Webster Reservoir) of surface area can be lost during typical dry conditions (lower fluctuation zone) while from 0.3% (Cheney Reservoir) to 20.8% (Kanopolis Reservoir) of area can be gained during typical wet conditions (upper fluctuation zone). The frequency, duration and spatial pattern of water level fluctuations in wetland ecosystems is an extremely important determinant of wetland structure and function. This water level flux is especially influential in determining wetland plant abundance, diversity and community composition (Bornette & Puijalon, 2011; Casanova & Brock, 2000; Van Geest, Coops, Roijackers, Buijse, & Scheffer, 2005). The amplitude and timing of water level fluctuations are critical to the establishment and maintenance of wetland plant communities, especially littoral communities such as those found in both large and small lakes (Deegan, White, & Ganf, 2007; Leira & Cantonati, 2008; Poff et al., 1997; Wilcox and Jerrine Nichols 2008). Water level variations impact the type of vegetation that is present in the wetland, which also affects the soils found there. Water depth, vegetation, and soil type are the

3.2. NWI The NWI was used for the identification of wetland habitats within each reservoir's flood pool boundary. NWI polygons were predominantly forested/shrub wetlands, followed by emergent wetlands, riverine wetlands, and lastly ponds, all of which are freshwater (Table 4). These four general classes are defined as follows: 1 Emergent – Wetlands characterized by rooted, erect hydrophytes (≈macrophytes), most of which are perennial species that are present throughout most of the year. These include all types of water regimes except subtidal and infrequently and irregularly exposed 20

Applied Geography 93 (2018) 16–24

K.M. Loeffler et al.

Fig. 4. 25th, 50th, and 75th percentile lake extents for (A) Elk City and (B) Webster reservoirs. Background is 2015 USDA-FSA National Agriculture Imagery Program data.

Table 3 Areal statistics for the 25th, 50th, and 75th percentile lake extents (ha) for the 20 federally owned study reservoirs in Kansas. Reservoir

Area of 25th percentile lake extent

Area of 50th percentile lake extent

Area of 75th Percentile Lake Extent

Area between 25th and 50th percentile lake extents

Area between 50th and 75th percentile lake extents

Percent area lost (50 → 25)

Percent area gained (50 → 75)

Big Hill Cheney Clinton Council Grove El Dorado Elk City Fall River Hillsdale John Redmond Kanopolis Kirwin Marion Melvern Milford Perry Pomona Toronto Tuttle Creek Webster Wilson Average

439 3953 2873 1097 2981 1317 914 1654 3634 1232 1244 2519 2465 5864 4028 1309 1018 4079 688 3424 2337

466 4076 2959 1156 3150 1426 932 1713 3789 1269 1865 2646 2565 6213 4142 1390 1055 4363 1254 3629 2503

470 4088 3156 1173 3180 1616 1004 1964 3995 1534 2077 2665 2705 6444 4603 1535 1137 5090 1471 3664 2679

27 123 86 59 169 109 18 59 155 37 621 126 100 349 114 81 37 285 566 205 166

4 11 197 17 30 190 72 251 206 264 212 19 140 231 461 145 81 727 216 36 176

5.8 3.0 2.9 5.1 5.4 7.6 2.0 3.4 4.1 2.9 33.3 4.8 3.9 5.6 2.8 5.9 3.5 6.5 45.1 5.6 8.0

0.9 0.3 6.6 1.5 0.9 13.3 7.7 14.7 5.4 20.8 11.4 0.7 5.5 3.7 11.1 10.4 7.7 16.7 17.2 1.0 8.0

21

Applied Geography 93 (2018) 16–24

K.M. Loeffler et al.

wetland cover pattern was also observed in the NWI wetland polygons that occur within the upper fluctuation zone. Based on the dominance of wetland categories there appear to be two groups of reservoirs: 1) reservoirs with predominantly forest/shrub wetlands, and 2) reservoirs with emergent wetlands that appear to be mostly shoreline communities (Table 4). It is unclear why the majority of these large reservoirs tend to be dominated by either emergent or forest/shrub wetland communities as they show no obvious corresponding pattern in ecoregion, basin shape, landscape, age or gradient/slope. Three reservoirs (Clinton, Elk City and Perry) did not fit well within the emergent and forest/shrub groupings. Clinton and Perry reservoirs tended to have a dominance of riverine wetlands (63 and 42%, respectively) with the second most common wetland category being emergent for Clinton and forest/shrub for Perry. Elk City Reservoir could have been included with the forest/ shrub grouping having just over 50% forest/shrub coverage but it was unique in that sub-dominant categories were riverine (27%) and pond (20%). Lastly, there were just four reservoirs that had very limited emergent wetland communities (i.e. shoreline communities); John Redmond, Kanopolis and Toronto reservoirs had less than one percent emergent wetland coverage while Pomona reservoir had less than five percent coverage.

Table 4 Percentages of types of NWI wetlands within the flood pool boundary (no parenthesis) and within the upper fluctuation zone (in parenthesis). The type with the maximum coverage within the upper fluctuation zone is bold. The primary NWI reservoir polygon was excluded from the calculations. Reservoir

Freshwater emergent

Freshwater forested/shrub

Freshwater pond

Riverine

Big Hill Cheney Clinton Council Grove El Dorado Elk City Fall River Hillsdale John Redmond Kanopolis Kirwin Marion Melvern Milford Perry Pomona Toronto Tuttle Creek Webster Wilson Average

7.1 (22.2) 14.3 (83.3) 61.9 (26.3) 23.0 (23.1) 9.1 (22.6) 6.3 (2.7) 7.4 (31.7) 3.4 (7.0) 15.2 (0.9) 9.3 (0.0) 39.9 (12.1) 21.8 (19.2) 41.1 (50.0) 17.9 (64.4) 46.9 (22.5) 24.0 (4.5) 5.7 (0.1) 35.7 (80.5) 35.0 (60.1) 83.3 (93.8) 25.4 (31.4)

81.4 (74.1) 37.7 (16.7) 6.9 (10.5) 74.3 (76.9) 89.9 (75.4) 30.9 (50.3) 70.0 (63.5) 95.3 (91.1) 69.7 (99.1) 55.2 (75.0) 59.6 (87.7) 76.7 (80.8) 31.5 (25.0) 14.9 (29.7) 34.6 (35.4) 48.9 (90.9) 86.6 (99.0) 22.3 (8.3) 58.6 (38.0) 6.2 (6.3) 52.6 (56.7)

11.5 (3.7) 0.8 (0.0) 2.6 (0.0) 0.5 (0.0) 0.3 (1.8) 16.0 (20.1) 2.8 (0.0) 1.3 (1.9) 3.4 (0.0) 2.1 (0.0) 0.5 (0.2) 1.5 (0.0) 6.4 (25.0) 0.6 (1.0) 7.8 (0.3) 1.9 (0.0) 7.1 (0.9) 2.1 (0.03) 2.7 (0.01) 9.7 (0.0) 4.1 (2.7)

0.0 (0.0) 47.2 (0.0) 27.0 (63.1) 2.1 (0.0) 0.8 (0.3) 46.9 (26.8) 19.8 (4.8) 0.0 (0.0) 11.7 (0.0) 33.5 (25.0) 0.1 (0.0) 0.0 (0.0) 21.0 (0.0) 66.5 (5.0) 10.8 (41.8) 25.2 (4.5) 0.6 (0.0) 39.9 (11.1) 3.7 (1.8) 0.8 (0.0) 17.9 (9.2)

3.3. Examining NWI propensity in the upper fluctuation zone using LiDAR slope As described earlier, we computed the median LiDAR slope for NWI features in the upper fluctuation zone and the upper flood pool, both of which were modified to exclude pixels with zero slope that almost invariably correspond with hydro-flattened features (and thus already identified as being a waterbody). For 14 of the 20 study reservoirs, median slope for NWI in the upper flood pool was greater than median slope for NWI in the upper fluctuation zone (Table 5), consistent with the hypothesis that relatively low-sloped ground in the upper fluctuation zone is more likely to be wetland than in the upper flood pool. On average across the 20 reservoirs, the median NWI slope from the upper flood zone captured 54.6% ( ± 11.7) of the NWI pixels in the upper fluctuation zone, with a maximum of 74.4% (Tuttle Creek) and a minimum of 32.0% (Marion). The average median percent slope for upper flood pool NWI was 3.2, whereas the average for the upper fluctuation zone was 2.8. Combining and analyzing the data from all 20 sites, the median percent slope for upper flood pool NWI was 2.1 compared to 1.4 for the upper fluctuation zone. In this collective assessment, applying the upper flood pool median NWI percent slope as a threshold captured 62.0% of the NWI area in the upper fluctuation zone, suggesting that relatively low-slope land in the upper fluctuation zone was 24% (=(62–50)/50) more likely to harbor NWI wetlands than similar land in the upper flood pool.

systems. The emergent vegetation adjacent to rivers and lakes is often referred to as "the shore zone" or the "zone of emergent vegetation" (Reid & Wood, 1976), and is most often considered separately from the river or lake. 2 Forest/shrub – A complex dominated by water-tolerant shrubs and trees typically located in the floodplain. Field inspection of palustrine forested and/or palustrine shrub wetlands occurring in our studies often found them to be dominated by willow (Salix spp.) and buttonbush (Cephalanthus occidentalis). Cowardin, Carter, Golet, and LaRoe (1979) lists two separate wetland classifications, forested wetland and scrub-shrub wetland; however, the NWI combines these two classifications. 3 Ponds – Typically small palustrine wetlands with an unconsolidated bottom or palustrine aquatic bed wetlands. Ponds can be natural or man-made and include farm ponds, stock ponds and vernal pools. 4 Riverine – Includes all wetlands and deepwater habitats contained within a channel, except for wetlands dominated by trees, shrubs, persistent emergents, emergent mosses, or lichens. This restrictive system excludes floodplains adjacent to the channel. A channel that conveys water all or some of the time or links two waterbodies can be a natural or artificially created feature. The NWI polygons were then restricted to the upper fluctuation zone. The largest wetland polygon within the upper fluctuation zone was a 15 ha freshwater forested/shrub wetland polygon above Kirwin Reservoir, and the smallest mapped wetland within this dataset was 0.1 acres, which was found in most of the mapped reservoirs describing different wetland types. The exclusion of NWI polygon areas that occurred outside the 75% high water zone resulted in two notable shifts within wetland categories. Both freshwater pond and riverine category percentages were almost reduced by one half, indicating that upland ponds and riverine areas comprise a larger fraction of wetland features in less frequently flooded regions associated with these reservoir systems. Likewise, if we consider only the upper fluctuation zone then the percent of all wetlands in the freshwater emergent and forest/shrub categories noticeably increases. The biggest increase was in the freshwater emergent category, which seems to comprise the “shoreline” wetland community as these polygons tend to follow the bathymetric/elevational contours. A

3.4. Summary and conclusion Hydrological characteristics of 20 federally operated reservoirs in Kansas were used to examine potential wetland development areas associated with the fluctuation zone of each reservoir, which we defined to be the area between the lake footprints given by the 25th and 75th water surface elevation percentiles using historic daily lake level data along with topographic information from LiDAR and bathymetry. The study reservoirs experienced an average area reduction of 8.0% between the 50th and 25th percentile lake extents, with a proportionally similar average area increase of 8.0% found between the 50th and 75th percentile lake extents. Though one might expect to see greater extent fluctuation in typical dry conditions in reservoirs with watersheds that receive less rainfall and where maintenance inflow is more of a challenge (or conversely, greater extent fluctuation in typical wet conditions in reservoirs with watersheds that generally provide sufficient or ample inflow), such patterns did not clearly bear out in the 22

Applied Geography 93 (2018) 16–24

K.M. Loeffler et al.

Table 5 Summary NWI-slope characteristics from the upper flood pool (UFP) and the upper fluctuation zone (UFZ). Areas with zero slope were excluded from the calculations to avoid already identified waterbodies. Area values are hectares and slope values are percents. Reservoir

Area UFP

Area UFZ

UFP \ UFZ as percent of flood zone

Area NWI in UFP

Area NWI in UFZ

Median NWI slope in UFP

Median NWI slope in UFZ

Percent NWI in UFZ with slope ≤ UFP NWI median slope

Big Hill Cheney Clinton Council Grove El Dorado Elk City Fall River Hillsdale John Redmond Kanopolis Kirwin Marion Melvern Milford Perry Pomona Toronto Tuttle Creek Webster Wilson sum or mean All combined

103 1076 2854 867 1026 3004 3228 1123 8285 4105 2126 1026 2710 6725 5575 2050 2848 16,559 1817 4410 71,517 71,516

4 11 197 17 30 190 72 251 206 264 212 19 140 231 461 145 81 727 216 36 3512 3512

96\4 99\1 94\6 98\2 97\3 94\6 98\2 82\18 98\2 94\6 91\9 98\2 95\5 97\3 92\2 93\3 97\3 96\4 89\11 99\1 95\5 95\5

17 211 227 95 198 276 372 118 2341 297 38 258 354 1145 1160 281 2344 2258 335 193 12,520 12,520

2 8 114 1 17 174 62 45 205 39 76 15 14 134 165 21 81 314 163 18 1670 1672

15.9 2.1 2.2 2.3 3.5 2.1 2.5 3.0 2.0 1.8 4.2 2.5 2.0 1.9 1.8 2.2 2.7 1.8 2.9 4.3 3.2 ± 3.1 2.1

11.2 3.9 1.7 1.5 5.8 1.3 1.2 1.2 1.6 2.7 1.8 5.1 2.5 1.8 1.0 0.9 2.4 0.6 1.7 6.0 2.8 ± 2.5 1.4

65.1 33.5 54.2 60.3 37.9 62.1 71.1 60.4 60.4 44.0 59.3 32.0 48.1 52.8 58.9 60.2 53.7 74.4 62.6 42.0 54.6 ± 11.7 62.0

to large-scale changes in wetted conditions above the study reservoirs. Other next steps include collecting water, sediment, and vegetation samples from the main basin and the upper fluctuation zones of these reservoirs to help determine whether the upper fluctuation zones are indeed developing into wetland ecosystems. Additionally, there is more to explore in the daily reservoir water level dataset used in this study, as the fluctuation of reservoir water level determines inundation frequency, duration, and depth in the upper fluctuation zone, all critical factors that affect wetland structure and function.

analysis. This could be a consequence of the small sample size coupled with the uniqueness of each individual reservoir situation (size, shape, inflow, etc.). All 20 reservoirs are highly managed with respect to water level and were constructed with the primary purpose of flood control. Using data from the USFWS National Wetland Inventory, we examined tendencies in wetland type distributions across the study reservoirs. Freshwater pond and riverine wetland classes comprised greater fractions of the NWI coverage in the flood pool above the 75th percentile extent (the “upper flood pool”) than between the 50th and 75th percentile extents (the “upper fluctuation zone”), consistent with greater resemblance of these drier portions of the flood pool with unimpounded portions of the stream. On the other hand, the increase in freshwater emergent and freshwater forested/shrub class fractions within the upper fluctuation zone is consistent with the increased hydrologic connectivity to the reservoir. By similar reasoning with regard to reservoir hydrologic connectivity and inundation frequency, one might generally expect lowsloped lands in the upper fluctuation zone to exhibit a greater propensity for wetland development than in the upper flood pool. To investigate, we examined NWI-slope statistics for the two regions using LiDAR. While the individual reservoir results are weakly consistent with this notion, a stronger result was obtained after combining all of the study sites. Specifically, we observed a 24% increase in NWI occurrence on low-sloped lands in the upper fluctuation zone compared to the upper flood pool, using median NWI slope from the upper flood pool as a maximum slope threshold. This assessment assumes completeness (or at least consistency) of wetland representation in the NWI across the reservoir flood zones. One possibility for further study would be to investigate remote sensing image archives to determine how these potential wetland development areas have changed over time. Though the present-day remote sensing data landscape continues to improve, the historical imagery that is readily available has a number of shortcomings, including lack of spatial resolution, temporal resolution, visibility, and contextual spatial information necessary to identify wetland status and condition or to generally capture the magnitude and frequency of typical wetland water level changes (Huggins, Jakubauskas, Baker, & Hammer, 2010). Nonetheless, a good place to start would be to examine the recently released, Landsat-based, long-term global surface water dataset (Pekel, Cottam, Gorelick, & Belward, 2016), which could shed light on medium

Acknowledgements This work was supported by the U.S. Environmental Protection Agency and Kansas Water Office [grant number CD 977513010CD97751901-1]. References deNoyelles, F., & Kastens, J. H. (2016). Reservoir sedimentation challenges Kansas. Transactions of the Kansas Academy of Science, 119(1), 69–81. Bornette, G., & Puijalon, S. (2011). Response of aquatic plants to abiotic factors: A review. Aquatic Sciences, 73(1), 1–14. Casanova, M. T., & Brock, M. A. (2000). How do depth, duration and frequency of flooding influence the establishment of wetland plant communities? Plant Ecology, 147(2), 237–250. Cooke, G. D., Welch, E. B., & Peterson, S. A. (2013). Lake and reservoir restoration. Ann Arbor, MI: Elsevier. Cooke, G. D., Welch, E. B., Peterson, S., & Nichols, S. A. (2016). Restoration and management of lakes and reservoirs. Boca Raton, FL: CRC press. Coops, H., Beklioglu, M., & Crisman, T. L. (2003). The role of water-level fluctuations in shallow lake ecosystems–workshop conclusions. Hydrobiologia, 506(1–3), 23–27. Coops, H., & Havens, K. E. (2005). Introduction. Hydrobiologia, 539(1), 169. Coppus, R., & Imeson, A. C. (2002). Extreme events controlling erosion and sediment transport in a semi-arid sub-Andean valley. Earth Surface Processes and Landforms, 27(13), 1365–1375. Cowardin, L. M., Carter, V., Golet, F. C., & LaRoe, E. T. (1979). Classification of wetlands and deepwater habitats of the United States. Washington, DC: Fish and Wildlife Service, US Department of the Interior. DASC (2016). Elevation data and LiDAR. Kansas Geological Survey, Kansas Data Access & Support Center, Accessed date: 12 April 2016. Deegan, B. M., White, S. D., & Ganf, G. G. (2007). The influence of water level fluctuations on the growth of four emergent macrophyte species. Aquatic Botany, 86(4), 309–315. Fox, G. A., Sheshukov, A., Cruse, R., Kolar, R. L., Guertault, L., Gesch, K. R., & Dutnell, R. C. (2016). Reservoir sedimentation and upstream sediment sources: Perspectives and future research needs on streambank and gully erosion. Environmental Management, 57(5), 945–955. García-Ruiz, J. M., Beguería, S., Nadal-Romero, E., González-Hidalgo, J., Lana-Renault, C.

23

Applied Geography 93 (2018) 16–24

K.M. Loeffler et al.

Peterson, D. L., Egbert, S. L., Price, K. P., & Martinko, E. A. (2004). Identifying historical and recent land-cover changes in Kansas using post-classification change detection techniques. Transactions of the Kansas Academy of Science, 107(3), 105–118. Poesen, J., Nachtergaele, J., Verstraeten, G., & Valentin, C. (2003). Gully erosion and environmental change: Importance and research needs. Catena, 50(2), 91–133. Poff, N. L., Allan, J. D., Bain, M. B., Karr, J. R., Prestegaard, K. L., Richter, B. D., Sparks, R. E., & Stromberg, J. C. (1997). The natural flow regime. BioScience, 47(11), 769–784. Rahmani, V., Hutchinson, S. L., Harrington, J. A., & Hutchinson, J. M. S. (2016). Analysis of frequency and magnitude of extreme rainfall events with potential impacts on flooding; a case study from the central United States. International Journal of Climatology, 36(10) 3578–358710.1002/joc.4577. Rahmani, V., Hutchinson, S. L., Harrington, J. A., Jr., Hutchinson, J. M. S., & Anandhi, A. (2015). Analysis of temporal and spatial distribution and change-points for annual precipitation in Kansas, USA. International Journal of Climatology, 35(13) 3879–388710.1002/joc.4252. Rahmani, V., Kastens, H. J., deNoyelles, F., Jakubauskas, E. M., Martinko, A. E., Huggins, H. D., Gnau, C., Liechti, M. P., Campbell, W. S., Callihan, A. R., et al. (2018). Examining storage capacity loss and sedimentation rate of large reservoirs in the central U.S. Great Plains. Water, 10, 190. Raje, D., & Mujumdar, P. P. (2010). Reservoir performance under uncertainty in hydrologic impacts of climate change. Advances in Water Resources, 33(3), 312–326. Reid, G. K., & Wood, R. D. (1976). Ecology of inland waters and estuaries (2nd ed.). New York City, NY: van Nostrand. Schenck, E., Wedel, K. L., & Monda, M. J. (1992). Classification of wetland and riparian areas in Kansas. Topeka, KS: Kansas Wetland & Riparian Areas Project. Schleiss, A. J., Franca, M. J., Juez, C., & De Cesare, G. (2016). Reservoir sedimentation. Journal of Hydraulic Research, 54(6), 595–614. Singh, H., Sinha, T., & Sankarasubramanian, A. (2014). Impacts of near-term climate change and population growth on within-year reservoir systems. Journal of Water Resources Planning and Management, 141(6), 04014078. Soballe, D. M., & Kimmel, B. L. (1987). A large-scale comparison of factors influencing phytoplankton abundance in rivers, lakes, and impoundments. Ecology, 68(6), 1943–1954. Soundharajan, B., Adeloye, A. J., & Remesan, R. (2016). Evaluating the variability in surface water reservoir planning characteristics during climate change impacts assessment. Journal of Hydrology, 538625–538639. Stene, E. O. (1946). How lakes came to Kansas. Transactions of the Kansas Academy of Science, 49, 117–137. Thornton, K. W., Kennedy, R. H., Carroll, J. H., Walker, W. W., Gunkel, R. C., & Ashby, S. (1981). Reservoir sedimentation and water quality—an heuristic model. Proceedings of the symposium on surface water impoundments (pp. 654–661). New York, NY: American Society of Civil Engineers. TWDB (2017). Glossary of terms. Texas Water Development Board. Available at: https:// waterdatafortexas.org/reservoirs/glossary, Accessed date: 30 January 2018. USACE (2013). Water supply storage reallocation John Redmond dam and reservoir, Kansas. Tulsa, OK: U.S. Army Corps of Engineers. USACE (2016a). United States Army Corps of Engineers. Kansas City District. Available at: http://www.nwk.usace.army.mil, Accessed date: 16 February 2016. USACE (2016b). United States Army Corps of Engineers. Tulsa District. Available at: http:// www.swt.usace.army.mil, Accessed date: 27 March 2016. USBR (2016). United States Bureau of Reclamation. United States Bureau of Reclamation. Available at: https://www.usbr.gov, Accessed date: 18 February 2016. USFWS (2016). National wetlands inventory. U.S. Fish and Wildlife Service. Available at: https://www.fws.gov/wetlands/, Accessed date: 28 January 2018. USGS (2016). United States Geological Survey. Kansas Office. Available at: https://www. usgs.gov, Accessed date: 19 January 2016. Valentin, C., Poesen, J., & Li, Y. (2005). Gully erosion: Impacts, factors and control. Catena, 63(2), 132–153. Verhoeven, J. T. A., Arheimer, B., Yin, C., & Hefting, M. M. (2006). Regional and global concerns over wetlands and water quality. Trends in Ecology & Evolution, 21(2), 96–103. Wilcox, D. A., & Jerrine Nichols, S. (2008). The effects of water-level fluctuations on vegetation in a Lake Huron wetland. Wetlands, 28(2), 487–501. Ziadat, F. M., & Taimeh, A. Y. (2013). Effect of rainfall intensity, slope, land use and antecedent soil moisture on soil erosion in an arid environment. Land Degradation & Development, 24(6) 582–59010.1002/ldr.2239.

N., & Sanjuán, Y. (2015). A meta-analysis of soil erosion rates across the world. Geomorphology, 239160–239173. Van Geest, G. J., Coops, H., Roijackers, R., Buijse, A. D., & Scheffer, M. (2005). Succession of aquatic vegetation driven by reduced water-level fluctuations in floodplain lakes. Journal of Applied Ecology, 42(2), 251–260. Goodin, D. G., Mitchell, J. E., Knapp, M. C., & Bivens, R. E. (1995). Climate and weather atlas of Kansas: An introduction. Educational ser. 12. Lawrence, KS: Kansas Geological Survey. Hargrove, W. (2008). Sedimentation in our reservoirs. Manhattan, KS: Kansas State University, Department of Communications. Hargrove, W., Johnson, D., Snethen, D., & Middendorf, J. (2010). From dust bowl to mud Bowl: Sedimentation, conservation measures, and the future of reservoirs. Journal of Soil and Water Conservation, 65(1), 17A. Hayes, N. M., Deemer, B. R., Corman, J. R., Razavi, N. R., & Strock, K. E. (2017). Key differences between lakes and reservoirs modify climate signals: A case for a new conceptual model. Limnology and Oceanography Letters, 2(2), 47–62. Heidemann, H. K. (2012). Lidar base specification. Reston, VA: United States Geological Survey. Huggins, D. G., Jakubauskas, M., Baker, D. S., & Hammer, T. (2010). Determining the utility and adaptability of remote sensing in monitoring and assessing reservoir eutrophication and turbidity for TMDL assessments. Lawrence, KS: Kansas Biological Survey. Johnson, D. (1992). Saving the wetlands from Agriculture: An examination of section 404 of the clean water Act and the conservation provisions of the 1985 and 1990 farm bills. Journal of Land Use & Environmental Law, 7(2), 299–319. Juracek, K. E. (2015). The aging of America's reservoirs: In-reservoir and downstream physical changes and habitat implications. JAWRA Journal of the American Water Resources Association, 51(1), 168–184. Keddy, P., & Fraser, L. H. (2000). Four general principles for the management and conservation of wetlands in large lakes: The role of water levels, nutrients, competitive hierarchies and centrifugal organization. Lakes and Reservoirs: Research and Management, 5(3), 177–185. Kennedy, R. H., Gunkel, R., Jr., & Thornton, K. W. (1982). The establishment of water quality gradients in reservoirs. Canadian Water Resources Journal, 7(1), 71–87. Kennedy, R. H., Thornton, K. W., & Ford, D. E. (1985). Characterization of the reservoir ecosystem. Microbial processes in reservoirs (pp. 27–38). Dordrecht, Netherlands: Springer. KGS (2017). Wetlands. Kansas Geological Survey. Available at: http://geokansas.ku.edu/ wetlands, Accessed date: 30 January 2018. Kimmel, B. L., & Groeger, A. W. (1984). Factors controlling primary production in lakes and reservoirs: A perspective. Lake and Reservoir Management, 1(1), 277–281. Kimmel, B. L., Lind, O. T., & Paulson, L. J. (1990). Reservoir primary production. Reservoir limnology: Ecological perspectives. New York City, NY: John Wiley & Sons. KWO (2015). John Redmond dredging initiative. Kansas Water Office. Available at: http:// kwo.ks.gov/projects/john-redmond-dredging, Accessed date: 13 September 2017. Lehner, B., Liermann, C. R., Revenga, C., Vörösmarty, C., Fekete, B., Crouzet, P., Döll, P., Endejan, M., Frenken, K., & Magome, J. (2011). High-resolution mapping of the world's reservoirs and dams for sustainable river-flow management. Frontiers in Ecology and the Environment, 9(9), 494–502. Leira, M., & Cantonati, M. (2008). Effects of water-level fluctuations on lakes: An annotated bibliography. Hydrobiologia, 613(1), 171–184. Lind, O. T. (1984). Patterns of phytoplankton populations and their relationship to trophic state in an elongate reservoir. SIL Proceedings, 1922–2010, 22(3), 1465–146910 1080/03680770.1983.11897519. Martinko, E., deNoyelles, J., Bosnak, K., Jakubauskas, M., Huggins, D., Kastens, J., Shreders, A., Baker, D., Blackwood, A., Campbell, S., & Rogers, C. (2014). Atlas of Kansas lakes: A resource for communities, policy makers and planners. Lawrence, KS: Kansas Biological Survey. Mouri, G., Shiiba, M., Hori, T., & Oki, T. (2011). Modeling reservoir sedimentation associated with an extreme flood and sediment flux in a mountainous granitoid catchment, Japan. Geomorphology, 125(2), 263–270. Nearing, M. A., Pruski, F. F., & O'neal, M. R. (2004). Expected climate change impacts on soil erosion rates: A review. Journal of Soil and Water Conservation, 59(1), 43–50. NOAA (2017). Glossary of hydrologic terms. National Weather Service, National Oceanic and Atmospheric Administration. Available at: http://www.nws.noaa.gov/om/hod/ SHManual/SHMan014_glossary.htm Accessed January 30, 2018 . Pekel, J., Cottam, A., Gorelick, N., & Belward, A. S. (2016). High-resolution mapping of global surface water and its long-term changes. Nature, 540(7633), 418.

24