landscapes

Aubrey Eckert-Gallup CE 598- Stream Restoration Final Term Paper

//Using the principles of landscape ecology to inform successful stream restoration projects. //
 * __ The Landscape Perspective __**


 * Abstract **

The principles of landscape ecology include the description of landscape components and their interconnected forms and processes based on the understanding of the fields of geography and biology. These principles can help to inform river restoration through the quantification of correlative effects between land use metrics and indicators of riverine health. Many case studies have found correlations between landscape factors and riverine health metrics such as water quality, hydrological form, and ecosystem function. Successful restoration projects can be designed following the determination of landscape components that contribute to degradation. Such projects must take into consideration the spatial and temporal scales over which landscape processes operate in order to understand the best possible actions for restoration. The application of a landscape perspective to river restoration will help to create and maintain sustainable natural river forms and functions for the benefit of river species, riparian vegetation, and future generations to come.


 * Introduction **

The approach to stream restoration has become increasingly focused on a systems dynamics perspective that utilizes an understanding of the collective processes present in a watershed in order to design a potentially successful restoration project. This perspective is related to the study of landscape ecology, which seeks to understand the quantification and evaluation of landscape components and to describe the connections between these components and their implications in the determination of ecological system health.

The systems approach to restoration must first be informed at its most basic level by an initial quantification of the various types of landscape components that are present in an area under study. The identification of the spatial variability that is present in a watershed at a given time or across a temporal span may help to inform the design of a restoration project by distinguishing areas of a landscape that might contribute to the project’s success or failure. This identification, however, may involve complex determinations that could be limited by the data and metrics used for quantification. Landscape component identification is further complicated by the importance of spatial scale to the dynamics of a system and their potential impacts on riverine health.

Following the quantification of the components present in a landscape under study, the dynamics of the system must be evaluated in order to understand how changes to any component in the system might impact the system as a whole. The change in these dynamics over time might also provide information regarding the manner in which the system came to be under stress and in need of restoration and, thus, may help to inform the design of a successful restoration project. Through the understanding of the dynamics of a complex landscape system, watershed managers may find that a restoration project in a component of the landscape away from the stream itself might create the most benefit to the system as a whole rather than a stream-based restoration focus.


 * The study of landscape ecology **

Landscape ecology evolved as a sub-discipline of geography synthesized with principles of hierarchy found in biology. An early definition of this amalgamated science describes landscape ecology as the characterization of a ‘landscape,’ a basic unit defined in geography, through the hierarchical presentation of its diverse components and functions (Saur, 1925). This early definition of landscape ecology also discusses the complexities of such a presentation and suggests the importance of the spatial scale of reference and the influence of the subjective nature of component characterization to a final presentation of landscape connectivity. The importance of these aspects of landscape characterization continue to be stressed in the field of landscape ecology today and will be discussed in their modern context in a subsequent section.

The continued evolution of the field of landscape ecology became heavily influenced by the principles of biology and the understanding of the hierarchical relationships among the organisms and forms present in an ecosystem. Watt (1947) describes the aggregation of patches of plant species within a defined community as the basic units that may be used in order to begin understanding the community’s structure and the dynamic relationships between each of its parts. The community as a whole is defined in this work as a “space-time mosaic” in which the temporal development of each species patch is influenced by and dependent on its surrounding neighbors. Watt justifies his conclusion of the importance of spatial relatedness through a study of seven plant species and the relationship between their life phases and spatial correlations. He concludes his work by citing the difficulty of extending such a characterization of correlation beyond the plant community to include the influence of other ecosystem components including animals and microorganisms, climate, and soils. The difficulty of the correct presentation of a landscape through a correlative understanding of its infinite parts serves as the basis for the modern understanding and practice of the science of landscape ecology.

Although the formulation and practice of the field of landscape ecology began in the early 1900’s, the definition of this science remained as an association between the fields of geography and biology with little description of its own foundational principles (Wiens et al., 2006). Though mid-century landscape ecologist Ernst Neef sought to define his field through a development of axiomatic principles (Neef, 1967), the true definition of landscape ecology as a scientific field of its own did not occur until the 1980’s. Risser et al. (1983) elicited a panel of 25 experts in order to form an overarching definition of the field of landscape ecology and to determine the potential applications of their discipline. This panel resulted in a final definition of landscape ecology as the study of the development, dynamics, and implications of spatial heterogeneity across both the abiotic and biotic components of landscapes through time with a further emphasis placed on the management of the phenomena associated with spatial heterogeneity.

The complex evolution of the field of landscape ecology mirrors the complexities of any study that might be undertaken under this discipline. The term ‘landscape’ itself implies the importance of spatial scale because an area of interest or importance must be determined as the first step in any study. An infinite number of components present in a chosen landscape area may be studied as a part of the correlated processes that make up the landscape as a whole. The outside processes (e.g., climate and man-made) that influence each of these landscape components and the related spatial and temporal scales of these processes impact the continued evolution of the landscape as a whole. The consideration of these complex components of landscape ecology and their relationship to the health of river systems is a critical step in the successful design of restoration projects.


 * Landscape ecology, land use, and the study of riverine systems **

The geological forms, ecological components, and driving processes of surrounding landscapes have a strong influence on the habitat and biological diversity of streams and rivers (Allan, 2004). Many case studies use the landscape perspective to inform their research by quantifying the correlations between landscape pattern and river process. The following sections describe the basic elements used for the study of the influence of landscape on elements of riverine health including methods for landscape classification and for the analysis of correlations between landscape and river health.


 * //Classifying landscape components //**

The initial determination and quantification of landscape types precedes the study of the spatial relationships between landscape components and the influence of these landscape components on river systems. The selection of a metric for determining landscape types may be influenced by the research goals of a specific study or the availability of data. As was previously stated, the division of a landscape into quantifiable patches or components can be formed through subjective assumptions that might influence the outcome of a study of correlation (Saur, 1925; Risser et al., 1983). Though some metrics for determining landscape types seek to avoid these subjective assumptions and their subsequent implications, others may be used due to their simplicity with the acknowledgement of the partially subjective nature of their methods.

Satellite imagery and geographic information systems (GIS) data can be used to quantify the components of a landscape under study (Bhat et al., 2006). These image-based data sources have become an increasingly important cost-effective and high resolution source of data for spatial assessment (Fernandes et al., 2010). Though these sources can provide a strong basis for landscape quantification, their use may be influenced by the subjective nature of the field knowledge used to delineate the discrete boundaries between any two component types (Buck et al., 2004). The effectiveness of the use of remote sensing data in landscape quantification is also influenced by pixel size and its relationship to the spatial scale of interest (Uuemaa et al., 2013). The data presented in these sources is then used to divide the landscape under consideration into pre-defined classes that can later be analyzed in order to determine their influence on a chosen indicator (Gergel et al., 2002).


 * //<span style="font-family: 'Arial','sans-serif';">Methods for correlating riverine health indicators with landscape components //**

<span style="font-family: 'Arial','sans-serif';">Following the delineation and quantification of landscape components, some method must be employed in order to assess the correlative relationships between these components and their impact on a river under study. The chosen method of analysis is driven by the indicator of riverine health of interest. These indicators include water quality, hydrologic metrics such as flow and channel shape, and the patterns in riparian vegetation of a river under consideration.

<span style="font-family: 'Arial','sans-serif';">Water quality and hydrologic metrics are often used as indicators in order to determine the influence of landscape and land use on the health of a river system. The analysis of this influence commonly begins with the collection of water quality data and the description of channel form, including bank stability and bed composition, of a study area. The impact of landscape and land use on these metrics can then be determined following the classification of the surrounding landscape. Statistical rank correlation and regression analyses can be used to determine the existence of correlations between these two data sets and can be used to determine which landscape components are the most important (Tran et al., 2010; Bhat et al., 2006, Wang et al., 1997). Following the calculation of correlation coefficients, scatterplots can be used to determine if any nonlinear relationships between the datasets exist that may have been missed (Buck et al., 2004).

<span style="font-family: 'Arial','sans-serif';">Riparian vegetation patterns are important indicators of riverine health because of their influence on ecological functions such as ecological structure, productivity, and diversity, their impact on water quality through water and sediment retention, and their important function of bank stabilization (Fernandes et al., 2010; Leyer et al., 2012). The influence of landscape pattern on riparian vegetation can be determined by the characterization of patterns in spatial variability. Spatial autocorrelation analyses and the application of a semivariogram function can also be used to determine the spatial independence between samples following the division of an area of interest into classes of importance. This allows for spatial autocorrelation to be removed so that regression analyses can be implemented in order to determine the impact of land use on riparian vegetation zones independently, without the additional correlation of riparian vegetation zones due to their spatial dependence on one another (Fernandes et al., 2010).

<span style="font-family: 'Arial','sans-serif';">Though correlation analyses can be successfully used to determine the impact of land use on riverine health, several challenges to this type of analysis must be considered. These challenges are described in depth by Allan (2004), as follows. First, the classification of landscape features as discrete states does not reflect the gradation of these features across an area of interest or the coexistence of multiple classification gradients within a single area. The coexistence of such gradients in reality can complicate the determination of influence through the tools discussed above. Additionally, nonlinear responses to landscape patterns can be found in indicator metrics, in part due to the gradient of land use as discussed previously. Further, the influence of legacy effects from disturbances that are no longer present must also be considered. Such disturbances might not be quantified in characterization of the landscape surrounding a system even if their effects are still present in the system under study. Finally, the impact of varying spatial scales is an important factor in the study of landscape and river health correlations. This factor will be discussed in the following section.


 * The impact of landscape on indicators of river health **

//<span style="font-family: 'Arial','sans-serif';">**Case study results** //

<span style="font-family: 'Arial','sans-serif';">Analyses correlating landscape patterns and riverine health commonly find that indicators of river health are correlated to some type of land use. All of the previously discussed indicators of river system health were found to be influenced by some aspect of land use within the case studies under consideration.

<span style="font-family: 'Arial','sans-serif';">Water quality measures were found to be strongly correlated with land use by Bhat et al. (2006), who found through regression analyses that all water quality parameters except for pH were explained by some aspect of the military use of the surrounding area. Buck et al. (2004) found that pastoral land use was strongly correlated to measured total nitrogen and turbidity, though other water quality parameters in their study were not found to be correlated with land use. Tran et al. (2010) found that land use within a defined buffer zone including urbanization and agricultural land use was strongly correlated with riparian ecology and stream channel structure although land use at the watershed scale did not have a strong influence on water quality indicators. Fernandes et al. (2010) found that agricultural land use close to the study area was correlated with structural changes in riparian vegetation while land use further from the study area had a lesser effect as has been found in other studies. Wang et al. (1997) discovered correlations between land use factors and physical indicators of river health including bank instability and sediment and bed composition.

<span style="font-family: 'Arial','sans-serif';">Reviews of numerous other case studies agree that the correlation between land use and metrics used as indicators of health is significant and well documented (Allan, 2004; Gergel et al., 2002). Agricultural land uses are often correlated with decreases in water quality and ecosystem structure as they act as sources of pollution including sediments, nutrients, and pesticides (Allan, 2004; Buck et al., 2004). Urban land use is strongly correlated with decreases in river health, often disproportionately so to the percent of total land cover that this land use classification comprises. The impacts of urban land use often include changes in hydrology due to increased runoff and decreased riparian areas, decreases in water quality due to pollution, and dramatic changes in channel and ecosystem structure (Tran et al., 2010; Allan, 2004). Thus, while varying land use classes are found to have differing impacts on indicators of river system health, it can be concluded that land use is strongly correlated with river system health and that anthropogenic changes to the natural processes present in river ecosystems often have a negative effect on the overall health of rivers in general.


 * //<span style="font-family: 'Arial','sans-serif';">Overarching conclusions //**

<span style="font-family: 'Arial','sans-serif';">The results of case studies and the discussion found in subject matter literature point to several important overarching factors relating landscape pattern and river health including the importance of spatial scale and the importance of temporal considerations.

<span style="font-family: 'Arial','sans-serif';">Spatial scale is an important factor both in the initial definition of a study area and its related processes and in the results of analyses correlating land use with impacts found in river health indicators. The influence of spatial scale in both of these areas in turn impacts the ability of research in this field to be transformed into the design of successful restoration projects. The processes that drive the conditions found in any river are the result of a hierarchical ordering of the ecosystem that the river is a part of. These processes may occur on a large spatial scale, such as climactic and geologic dynamics, or on a small spatial scale, such as chemical interactions and reactions. The understanding of the varying spatial scales of processes that are important to the natural flux of a river system is a critical step in the determination of impacts caused by landscape processes on indicators of riverine health (Allan, 2004; Allan, 1997; Frissell et al., 1986).

<span style="font-family: 'Arial','sans-serif';">Several studies have been designed in order to specifically discern the relationship between the proximity and spatial scale of land use and the impact that land use has on indicators of river system health. Tran et al. (2010) structured their study area with a 200 meter buffer on either side of a stream, characterizing the zone of near-field land use. The impact of this land use was compared to that of the far-field study area outside of the defined buffer zone. This study concluded that the near-field landscape and land use was more strongly correlated to water quality indicators than the factors governing far-field land use, supporting the idea that the discussion of spatial scale is an important component of any study of this type. Zhou et al. (2012) analyzed the influence of landscape factors on river flow and water quality parameters at the subwatershed, catchment, and buffer (500 m, 1000 m, 1500 m, and 5000 m) spatial scales. Their analyses found that correlations between landscape patterns and river health indicators varied across their specified spatial scales. In contrast to the results of Tran et al. (2010), Zhou et al. (2012) found that the effect of landscape pattern on water quality variables was more important at larger spatial scales than at small spatial scales. The conflicting results of these two studies again raise the importance of the consideration of the subjective nature of certain components of landscape ecology. Though these two studies considered many of the same river health indicators, their study areas may be influenced by different natural and anthropogenic processes. These differences may manifest themselves at varying spatial scales, contributing to contrasting findings from these similarly structured studies.

<span style="font-family: 'Arial','sans-serif';">The consideration of the temporal aspects present in both the quantification of landscape patterns and the measurement of river health indicators is an important, but sometimes disregarded, element in the study of the impact of landscape ecology on river systems. Many of the studies reviewed in this paper considered only the landscape patterns found in their study area at a certain point in time (Tran et al., 2010; Zhou et al., 2012; Bhat et al., 2006; Fernandes et al., 2010). This consideration does not allow for the recognition of the effects of past land uses that may no longer be visible on the landscape but might still impact processes that influence river health indicators over longer temporal scales, a possible pitfall of this type of analysis that was discussed by Allan (2004). While the inclusion of temporal considerations increases the complexity of analysis, these considerations may help to create a more meaningful and rigorous correlative result. Buck et al. (2004) studied the influence of short-scale temporal change by examining correlations between the rate of stocking (a time-dependent variable) and water quality parameters through a regression analysis. This study also considered the temporal changes in water quality with the intent of informing management practices in the study area. The temporal considerations found in this example, though limited, suggest possible extensions of the work reviewed in other study areas.


 * Using the landscape perspective to inform river restoration **

<span style="font-family: 'Arial','sans-serif';">River restoration should be informed by the correlation found between landscape patterns and land use with indicators of river health. The negative impacts that landscape elements have on river health can only be addressed by the understanding of their driving processes as opposed to simply restoring their original form. This type of process-based restoration will ultimately lead to more positive, long-term restoration results.


 * //<span style="font-family: 'Arial','sans-serif';">Motivating Principles //**

<span style="font-family: 'Arial','sans-serif';">The characterization of the processes driving river degradation can be completed through the correlation of river health indicators with land use and ecosystem processes as was described previously. This allows for the identification of the causes of degradation, the first step towards the creation of process-based restoration (Beechie et al., 2010). Such an approach avoids the pitfalls of form-based restoration by addressing the root of the issue at hand; instead of simply recreating the original form of the river in a project that will ultimately fall prey to the same destructive affects, the source of the degradation can be corrected so that the river can return to equilibrium (Beechie et al., 2010; Roni et al., 2002). The recognition of the natural processes governing the river system and the characterization of both their spatial and temporal scales of importance is also an important step in this phase of process-based restoration as this understanding such system dynamics will contribute to the success of a restoration project by allowing for the integration of natural processes into project design (Benda et al., 2011). Thus, the first step towards any process-based restoration project must be the identification of the processes themselves, both man-made and natural, which may include importance and correlation analyses such as those discussed earlier in this work. A diagram showing the linkages between landscape controls, processes, and habitat effects from Roni et al. (2002) is shown in Figure 1 below. // Figure 1: From Roni et al. (2002) – schematic diagram showing the linkages between landscape controls, processes, and habitat effects. //

<span style="font-family: 'Arial','sans-serif';">The scale of any restoration project must be matched to the scale of the driving processes in order for the restoration project to succeed (Beechie et al., 2010). The importance of spatial scale in understanding the processes that affect indicators of river health was an overarching conclusion of the research reviewed earlier in this work. The attempt to promote restoration through actions against a single source of degradation within a watershed while other sources of the same type of degradation remain unaddressed throughout the watershed will not create the desired restorative outcome. In order to complete such a project, the spatial scale of the processes at work must be quantified and a restoration project must be designed to match this scale (Benda et al., 2011; Beechie et al., 2010). A visual description of watershed processes and their related spatial and temporal scales from Beechie et al. (2010) is shown in Figure 2 below.

// Figure 2: From Beechie et al. (2010) - visual description of watershed processes and their related spatial and temporal scales. //

<span style="font-family: 'Arial','sans-serif';">The temporal scale of processes driving the health of a system under consideration must also be determined in order to contribute to the success of restoration. As was discussed previously, indicators of river health may be responding to historical sources of degradation that are not visible in the landscape because the processes affected by these sources might only show the impact of degradation over a long temporal scale (Allan, 2004). These temporal scales also impact the time over which the results of restoration may be expected to become apparent. The results of restoration of processes that cause degradation over a long period of time may only become apparent over a proportionate period of time (Beechie et al., 2010). The understanding of the temporal processes at work in a system under consideration will thus help to inform both restoration actions and their expected outcomes.


 * //<span style="font-family: 'Arial','sans-serif';">Description of landscape and process based restoration actions //**

<span style="font-family: 'Arial','sans-serif';">Reviews of landscape and process based river restoration actions generally separate projects into channel-centric actions and actions that focus on areas further away from the channel. Chanel-centric actions are those that focus on restoring processes or habitat that occur directly in the stream or river channel. These actions can, in turn, have restorative impacts on the landscape and ecosystems surrounding the river channel. Actions that focus on areas away from the channel improve components of ecosystem processes that impact indicators of river health. It is important to note that, while restoration through habitat creation is not by definition an explicitly process-based restoration action, this more form-based work can serve to restore many related ecosystem processes and functions. Thus, habitat creation both in and around a stream channel is often included in reviews of landscape related restoration actions.

<span style="font-family: 'Arial','sans-serif';">Channel-centric actions described in reviews of process-based restoration practices include efforts to restore the connectivity of riverine ecosystems with the purpose of restoring habitat processes and function both directly in the channel and in its surrounding areas. The introduction of environmental flow regimes that follow important components of a natural hydrograph in heavily managed rivers helps to restore channel processes that govern the function of many different ecosystem components (Beechie et al., 2010). The creation of fish passages through artificial barriers is another example of a restoration action that seeks to restore the processes inherent in ecosystem connectivity within a river system. Such actions can greatly impact the availability of habitat for and productivity of in-stream species such as fish and near-stream species and vegetation (Roni et al., 2002). The addition of natural debris into a river system may be used to both create riverine habitat and to restore processes that impact river form including the formation of pools or the reintroduction of overbank flooding (Roni et al., 2002; Beechie et al., 2010). The restoration of and creation of estuarine habitats including the removal of connectivity barriers, sediment excavation, and other related actions have been found to restore the habitat required for the life cycle needs of fish and other species. Finally, the restoration of natural nutrient availability through the placement of fish carcasses in rivers in the Pacific Northwest has been used to augment the nutrient deficiencies created by decreasing salmon populations that impact in-stream and near-stream organisms and vegetation (Roni et al., 2002). Thus, channel-centric restoration actions informed by interactions between the river and its surrounding landscape promote restoration success by recognizing that the restoration of the interconnected network of ecosystem processes can impact a wide variety of river health indicators.

<span style="font-family: 'Arial','sans-serif';">Restoration actions focusing on areas away from the channel include habitat and riparian restoration actions and changes to land management practices. Riparian-based restoration actions focus on implementing changes to riparian form in order to restore related river processes. The removal of overstory or understory vegetation in order to encourage the growth of tree or vegetation species of importance to a river process of interest is an example of such a restoration action (Roni et al., 2002). Seedling planting along with a decrease in competitive species is often used as a measure for riparian restoration. These actions are often taken in order to meet the goal of restoring indicators of water quality (Emmingham et al., 2000). Though such actions have been widely implemented, Emmingham et al. (2000) notes that these types of restoration practices must be coupled with a longer term management strategy in order to achieve success. In addition to actions aimed at the restoration of riparian areas and their related processes, changes in land management are another example of restoration efforts that focus on areas away from a river channel. Restoration through the introduction of changes in land management practices including the reduction of grazing and other livestock related uses can help to restore a variety of river processes including channel form and water quality (Roni et al., 2002). Thus, restoration actions that are not channel-centric can strongly impact indicators of river health, especially when such actions are applied and supplemented over a long period of time. The possibilities for the success of such restoration actions are supported by the many previously mentioned studies that found correlations between land use and land management and indicators of river health.


 * Conclusion **

<span style="font-family: 'Arial','sans-serif';">The use of a landscape systems dynamics perspective to both quantify the sources of river degradation and to inform the design of restoration projects is supported by many case studies and reviews of restoration practices. The development of an understanding of the collective processes present in a watershed including the characterization of the manner in which landscape forms and functions impact indicators of river health is a complex and valuable first step towards any type of restoration decision. The application of the founding geographic and biological principles of landscape ecology to watershed evaluation and restoration inherently requires the recognition of varying spatial and temporal scales of importance. Though such works present huge challenges to both restoration design and sustainable management, the invaluable nature of their use cannot be understated because of the possibilities that they create for long term success. The realization of river restoration design goals can only help to promote the importance of regaining the natural form and function of our rivers and streams for the benefit of future generations.


 * References **

<span style="color: #262626; font-family: 'Arial','sans-serif';">Allan, J. D. 2004. Landscapes and Riverscapes: The Influence of Land Use on Stream Ecosystems. Annual Review of Ecology, Evolution & Systematics 35, no. 1: 257-284.

<span style="color: #262626; font-family: 'Arial','sans-serif';">Allan, J. D., D. L. Erickson, and J. Fay. 1997. The influence of catchment land use on stream integrity across multiple spatial scales. Freshwater Biology 37: 149-161.

<span style="color: #262626; font-family: 'Arial','sans-serif';">Beechie et al. 2010. Process-based Principles for Restoring River Ecosystems. Bioscience 60, no. 3: 209-222.

<span style="color: #262626; font-family: 'Arial','sans-serif';">Benda, L., B. <span style="font-family: 'Arial','sans-serif';">Miller and J. Barquin. 2011. Creating a catchment perspective for river restoration. Hydrology & Earth System Sciences Discussions 8, no. 2: 2929-2973.

<span style="color: #262626; font-family: 'Arial','sans-serif';">Bhat S. et al. 2006. Relationships between stream water chemistry and military land use in forested watersheds in Fort Benning, Georgia. Ecological Indicators 6: 458-466.

<span style="color: #262626; font-family: 'Arial','sans-serif';">Buck, O., D. Niyogi, and C. Townsend. 2004. Scale-dependence of land use effects on water quality of streams in agricultural catchments. Environmental Pollution 130: 287-299.

<span style="color: #262626; font-family: 'Arial','sans-serif';">Emmingham et al. 2000. Silviculture Practices for Riparian Forests in the Oregon Coast Range. Oregon State University, Forest Research Laboratory, Research Contribution 24: 1-38.

<span style="color: #262626; font-family: 'Arial','sans-serif';">Fernandes, M. R., F. C. Aguiar, and M. T. Ferreira. 2011. Assessing riparian vegetation structure and the influence of land use using landscape metrics and geostatistical tools. Landscape & Urban Planning 99, no. 2: 166-177.

<span style="color: #262626; font-family: 'Arial','sans-serif';">Frissell, C. et al. 1986. A hierarchical framework for stream habitat classification: Viewing streams in a watershed context. Environmental Management 10, no. 2: 199-214.

<span style="color: #262626; font-family: 'Arial','sans-serif';">Gergel et al. 2002. Landscape indicators of human impacts to riverine systems. Aquatic Sciences 64, no. 2: 118-128.

<span style="color: #262626; font-family: 'Arial','sans-serif';">Leyer I., et al. E. Mosner, and B. Lehmann. 2012. Managing floodplain-forest restoration in European river landscapes combining ecological and flood protection issues. Ecological Applications 22, no. 1: 240-249.

<span style="color: #262626; font-family: 'Arial','sans-serif';">Neef, E. 1967. The theoretical foundations of landscape study. Translated by Olaf Bastian. Geographisch-Kartographische, Ansalt ed. 18-38. Gotha/Leipzig: VEB Hermann Haack. Rpt. in Foundation Papers in Landscape Ecology. Ed. Wiens et al. New York: Columbia University Press, 2006. 225-245.

<span style="color: #262626; font-family: 'Arial','sans-serif';">Risser, P., J. Karr and R. Forman. 1983. Landscape ecology: directions and approaches. Illinois Natural History Survey Special Publication no. 2. Champaign: Illinois Natural History Survey. Rpt. in Foundation Papers in Landscape Ecology. Ed. Wiens et al. New York: Columbia University Press, 2006. 254-264.

<span style="color: #262626; font-family: 'Arial','sans-serif';">Roni et al. 2002. A review of stream restoration techniques and a hierarchical strategy for prioritizing restoration in Pacific Northwest watersheds. North American Journal of Fisheries Management 22, no. 1: 1-20.

<span style="color: #262626; font-family: 'Arial','sans-serif';">Saur, C. 1925. The morphology of landscape. University of California Publications in Geography 2: 19-53. Rpt. in Foundation Papers in Landscape Ecology. Ed. Wiens et al. New York: Columbia University Press, 2006. 36-70.

<span style="color: #262626; font-family: 'Arial','sans-serif';">Tran et al. 2010. Land-use proximity as a basis for assessing stream water quality in New York State (USA). Ecological Indicators 10: 727-733.

<span style="color: #262626; font-family: 'Arial','sans-serif';">Uuemaa, E., U. Mander, and R. Marja. 2013. Trends in the use of landscape spatial metrics as landscape indicators: A review. Ecological Indicators 28, 100-106.

<span style="color: #262626; font-family: 'Arial','sans-serif';">Watt, A. S. 1947. Pattern and process in the plant community. Journal of Ecology 35: nos. 1 and 2: 1-22. Rpt. in in Foundation Papers in Landscape Ecology. Ed. Wiens et al. New York: Columbia University Press, 2006. 102-123.

<span style="color: #262626; font-family: 'Arial','sans-serif';">Wang et al. 1997. Influences of Watershed Land Use on Habitat Quality and Biotic Integrity in Wisconsin Streams. Fisheries 22, no.6: 6-12.

<span style="color: #262626; font-family: 'Arial','sans-serif';">Wiens et al. 2006. Foundation papers in landscape ecology. New York: Columbia University Press.

<span style="color: #262626; font-family: 'Arial','sans-serif';">Zhou, T., J.G. Wu, and S.L. Peng. 2012. Assessing the effects of landscape pattern on river water quality at multiple scales: A case study of the Dongjiang River watershed, China. Ecological Indicators 23: 166-175.