Drivers Of Land Use Change Photos
Executive SummaryClimate can affect and be affected by changes in land cover (the physical features that cover the land such as trees or pavement) and land use (human management and activities on land, such as mining or recreation). A forest, for instance, would likely include tree cover but could also include areas of recent tree removals currently covered by open grass areas. Land cover and use are inherently coupled: changes in land-use practices can change land cover, and land cover enables specific land uses.
Understanding how land cover, use, condition, and management vary in space and time is challenging.Changes in land cover can occur in response to both human and climate drivers. For example, demand for new settlements often results in the permanent loss of natural and working lands, which can result in localized changes in weather patterns, temperature, and precipitation.
Aggregated over large areas, these changes have the potential to influence Earth’s climate by altering regional and global circulation patterns, changing the albedo (reflectivity) of Earth’s surface, and changing the amount of carbon dioxide (CO 2) in the atmosphere. Conversely, climate change can also influence land cover, resulting in a loss of forest cover from climate-related increases in disturbances, the expansion of woody vegetation into grasslands, and the loss of beaches due to coastal erosion amplified by rises in sea level.Land use is also changed by both human and climate drivers. Land-use decisions are traditionally based on short-term economic factors. Land-use changes are increasingly being influenced by distant forces due to the globalization of many markets. Land use can also change due to local, state, and national policies, such as programs designed to remove cultivation from highly erodible land to mitigate degradation, legislation to address sea level rise in local comprehensive plans, or policies that reduce the rate of timber harvest on federal lands. Technological innovation has also influenced land-use change, with the expansion of cultivated lands from the development of irrigation technologies and, more recently, decreases in demand for agricultural land due to increases in crop productivity. The recent expansion of oil and gas extraction activities throughout large areas of the United States demonstrates how policy, economics, and technology can collectively influence and change land use and land cover.Decisions about land use, cover, and management can help determine society’s ability to mitigate and adapt to climate change.
The figure shows the net change in land cover by class in square miles, from 1973 to 2011. Land-cover change has been highly dynamic over space, time, and sector, in response to a range of driving forces. Net change in land cover reveals the trajectory of a class over time. A dramatic example illustrated here is the large decline in agricultural lands in the two Great Plains regions beginning in the mid-1980s, which resulted in large part from the establishment of the Conservation Reserve Program.
Over the same period, agriculture also declined in the Southwest region; however, the net decline was largely attributable to prolonged drought conditions, as opposed to changes in federal policy. Data for the period 1973–2000 are from Sleeter et al. 2013, while data from 2001–2011 are from the National Land Cover Database (NLCD). Note: the two disturbance categories used for the 1973–2000 data were not included in the NLCD data for 2001–2011 and largely represent conversions associated with harvest activities (mechanical disturbance) and wildfire (nonmechanical disturbance).
Comparable data are unavailable for the U.S. Caribbean, Alaska, and Hawai’i & U.S.-Affiliated Pacific Islands regions, precluding their representation in this figure. From Figure 5.2 (Source: USGS). Climate can affect and be affected by changes in land cover (the physical features that cover the land, such as trees or pavement) and land use (human management and activities on land, such as mining or recreation). A forest, for instance, would likely include tree cover but could also include areas of recent tree removals currently covered by open grass areas.
Land cover and use are inherently coupled: changes in land-use practices can change land cover, and land cover enables specific land uses. Understanding how land cover, use, condition, and management vary in space and time is challenging, because while land cover and condition can be estimated using remote sensing techniques, land use and management typically require more local information, such as field inventories. Identifying, quantifying, and comparing estimates of land use and land cover are further complicated by factors such as consistency and the correct application of terminology and definitions, time, scale, data sources, and methods. While each approach may produce land-use or land-cover classifications, each method may provide different types of information at various scales, so choosing appropriate data sources and clearly defining what is being measured and reported are essential.Changes in land cover can occur in response to both human and climate drivers.
For example, the demand for new settlements often results in the permanent loss of natural and working lands, which can result in localized changes in weather patterns, temperature, and precipitation. Humans have had a far-reaching impact on land cover within the contiguous United States. Of the approximately 3.1 million square miles of land area, approximately 28% has been significantly altered by humans for use as cultivated cropland and pastures (22%) or settlements (6%; Figure 5.1a). Land uses associated with resource production (such as grazing, cropland, timber production, and mining) account for more than half of the land area of the contiguous United States, followed by land that is conserved (16%), built-up areas (13%), and recreational land (10%; Figure 5.1b). Between 2001 and 2011, developed land cover increased by 5% and agriculture declined by 1%. Urbanization was greater between 2001 and 2006 than between 2006 and 2011, which may be attributable to the 2007–2009 economic recession., The relative stability in agricultural land use between 2001 and 2011 masks widespread fluctuations brought about by the abandonment and expansion of agricultural lands (see Figure 5.2 for more detail). Figure 5.1: The composition of land use and land cover (LULC) is highly variable across the United States, owing in part to the natural environmental settings of each region.
Forests dominate much of the vegetated areas of the eastern United States, while much of the Great Plains and Southwest are dominated by grasses and shrubs. Characterizing the composition of LULC also depends on the type of classification system used. NCA RegionCroplandsForestlandsGrasslandsOther LandsSettlementsWetlandsAlaska105,5864,336Hawai'i1732,5011,991Midwest212,93,7534,8,867Northern Great Plains136,48,6784,4738,2169,765Northeast24,41,6492,2,521Northwest28,09,9633,8537,7845,573Southern Great Plains103,682,2162,790Southeast84,18,4423,4,852Southwest39,716,2,31110,237Total629,5501,166,3381,358,866,478163,992. Table 5.1: Definitions of land use and land cover vary among agencies and entities collecting those data. This may lead to fundamental differences in these estimates that must be considered when comparing estimates of cover and use. For the purposes of this report, land cover is defined as the physical characteristics of land, such as trees or pavement, and land use is characterized by human management and activities on land, such as mining or recreation.
The land-use area estimates in this table and throughout this chapter were obtained from the U.S. Forest Service’s Forest Inventory and Analysis (FIA) Program and the National Resources Conservation Service’s (NRCS) Natural Resources Inventory (NRI) data, when available for an area, because the surveys contain additional information on management, site conditions, crop types, biometric measurements, and other data that are needed to estimate carbon stock changes and nitrous oxide and methane emissions on those lands. If NRI and FIA data are not available for an area, however, then the NLCD product is used to represent the land use. Since all three data sources were used in the land representation analysis within the National Inventory Report, we used land-use estimates from the U.S. Environmental Protection Agency’s annual greenhouse gas inventory report.
Data are unavailable for both the U.S. Caribbean region and the U.S.-Affiliated Pacific Islands in the NRI and FIA datasets. Vegetated land cover, including grasslands, shrublands, forests, and wetlands, accounted for approximately two-thirds of the contiguous U.S. Land area and experienced a net decline of approximately 5,150 square miles between 2001 and 2011. However, many of these areas are also used for the production of ecosystem goods and services, such as timber and grazing, which lead to changes in land cover but may not necessarily result in a land-use change.
Between 2001 and 2011, forest land cover had the largest net decline of any class (25,730 square miles) but forest land use increased by an estimated 3,200 square miles over a similar period. The increase in forest land use is due, in large part, to the conversion of abandoned croplands to forestland and the reversion to and expansion of trees in grassland ecosystems in the Great Plains and western United States. There have also been losses in forest land use over the past 25 years, predominantly to grasslands and settlements, with grasslands and shrublands increasing in area by nearly 20,460 square miles. Collectively, non-vegetated areas, including water, barren areas, and snow and ice, account for approximately 6% of the total land area. Figure 5.2: The figure shows the net change in land cover by class in square miles, from 1973 to 2011. Land-cover change has been highly dynamic over space, time, and sector, in response to a range of driving forces.
Net change in land cover reveals the trajectory of a class over time. A dramatic example illustrated here is the large decline in agricultural lands in the two Great Plains regions beginning in the mid-1980s, which resulted in large part from the establishment of the Conservation Reserve Program. Over the same period, agriculture also declined in the Southwest region; however, the net decline was largely attributable to prolonged drought conditions, as opposed to changes in federal policy. Data for the period 1973–2000 are from Sleeter et al.
(2013), while data from 2001–2011 are from the National Land Cover Database (NLCD). Note: the two disturbance categories used for the 1973–2000 data were not included in the NLCD data for 2001–2011 and largely represent conversions associated with harvest activities (mechanical disturbance) and wildfire (nonmechanical disturbance).
Comparable data are unavailable for the U.S. Caribbean, Alaska, and Hawai’i & U.S.-Affiliated Pacific Islands regions, precluding their representation in this figure. Source: USGS. Coastal regions, as mapped within the National Oceanic and Atmospheric Administration’s (NOAA) Coastal Change Analysis Program (C-CAP), account for 23% of the contiguous U.S. Land area and have been particularly dynamic in terms of change, accounting for approximately 50% of all land-cover change and 43% of all urbanization in the contiguous United States.
Approximately 8% of the coastal region changed between 1996 and 2010, which included about 16,500 square miles of forest loss and about 5,700 square miles of gain in urban land, a rate three times higher than that of the interior of the United States. Additionally, nearly 1,550 square miles of wetlands were lost in coastal regions, a trend counter to that of the Nation as a whole. A majority of this wetland loss has occurred in the northern Gulf of Mexico (; ). Coastal shoreline counties comprise approximately 10% of the United States in terms of land cover (excluding Alaska and the U.S. Caribbean) yet represent 39% of the U.S.
Population (2010 estimates), with population densities six times higher than in non-coastal areas. Between 1970 and 2010, the population in coastal areas increased by nearly 40% and is projected to increase by an additional 10 million people over 2010–2020 (Figure 5.3). Increases in the frequency of high tide flooding and extreme weather events (such as hurricanes and nor’easters), wetland loss, and beach loss from sea level rise present potential threats to people and property in the coastal zone (;; ). Figure 5.3: The figure shows the development-related changes surrounding Houston, Texas, from 1996 to 2010, as mapped by NOAA’s Coastal Change and Analysis Program (C-CAP). Areas of change between 1996 and 2010 are shown in black.
These changes can have numerous impacts on the environment and populations, ranging from increased urban heat island effects and storm water runoff (the latter of which can increase flooding and produce water quality impacts), to decreases in natural cover. Source: USGS. Disturbance events (such as wildfire and timber harvest) are important factors that influence land cover. For example, forest disturbances can initiate a succession from forest to herbaceous grasslands to shrublands before forest reestablishment, with each successional stage having a different set of feedbacks with the climate. The length of an entire successional stage varies based on local environmental characteristics.
Permanent transitions to new cover types after a disturbance are also possible for many reasons, including the establishment of invasive or introduced species that are able to quickly establish and outcompete native vegetation., Data from the North American Forest Dynamics dataset indicate that forest disturbances affected an average of approximately 11,200 square miles per year in the contiguous United States from 1985 to 2010 (an area greater than the entire state of Massachusetts). Between 2006 and 2010, the rate of forest disturbance declined by about one-third. Although these data include a wide range of disturbance agents, including fire, insects, storms, and harvest, the sharp decline likely corresponds to a reduction in timber harvest activities resulting from a drop in demand for construction materials following the 2007–2009 economic recession.Wildland fires provide a good example of how ecosystem disturbance, climate change, and land management can interact.
Between 1979 and 2013, the number of days with weather conditions conducive to fire has increased globally, including in the United States. At the same time, human activities have expanded into areas of uninhabited forests, shrublands, and grasslands, exposing these human activities to greater risk of property and life loss at this wildland–urban interface., Over the last two decades, the amount of forest area burned and the expansion of human activity into forests and other wildland areas have increased. These changes in climate and patterns of human activity have led in part to the development of a national strategy for wildland fire management for the United States. The strategy, published in 2014, was one outcome of the Federal Land Assistance, Management, and Enhancement (FLAME) Act of 2009. An important component of the national strategy is a classification of U.S. Counties based on their geographic context; fire history; amount of urban, forest, and range land; and other factors.
Causes Of Land Use Change
The land-use, land-cover, and other components of the classification model are used to guide management actions. Future ChangesRepresentative Concentration Pathways (RCPs) were developed to improve society’s understanding of plausible climate and socioeconomic futures. Climate can drive changes in land cover and land use in several ways, including changes in the suitability of agriculture , increases in fire frequency and extent , the loss or migration of coastal wetlands, and the spatial relocation of natural vegetation.
Effects Of Land Use Change
The extent of the climate influence is often difficult to determine, given that changes occur within interconnected physical and socioeconomic systems, and there is a lack of comprehensive observational evidence to support the development of predictive models, leaving a large degree of uncertainty related to these future changes. Models can be used to demonstrate how climate change may impact the production of a given agricultural commodity and/or suggest a change in land use (for example, econometric models, global gridded crop models, and integrated assessment models). However, the true impact may be mitigated by the influence of global economic markets, a shift to a different crop that is better suited to the new climate pattern, technological innovations, policy incentives, or capital improvement projects.
This area of integrated, multidisciplinary scientific research is just emerging.Important feedbacks with agriculture are anticipated under changing climate conditions. Recent trends show a shift from dryland farming to irrigated agriculture throughout much of the Great Plains region (; ). Future projections suggest that cropland suitability may increase at higher latitudes and that croplands could shift to livestock grazing southward. For high-latitude regions, climate change could result in a large-scale transformation from naturally vegetated ecosystems to agronomy-dominated systems. Climate warming also could result in a shift from higher-productivity systems (such as irrigated agriculture) to lower-productivity systems (such as dryland farming).
Due to the globally interconnected nature of agricultural systems, climate change has broad implications for food security. Process DescriptionChapter authors developed the chapter through technical discussions, literature review, and expert deliberation via email and phone discussions. The authors considered feedback from the general public, the National Academies of Sciences, Engineering, and Medicine, and federal agencies. For additional information about the overall process for developing the report, see.The topic of land-use and land-cover change (LULCC) overlaps with numerous other national sectoral chapters (for example,;; ) and is a fundamental characteristic of all regional chapters in this National Climate Assessment.
This national sectoral chapter thus focuses on the dynamic interactions between land change and the climate system. The primary focus is to review our current understanding of land change and climate interactions by examining how land change drives changes in local- to global-scale weather and climate and how, in turn, the climate drives changes in land cover and land use through both biophysical and socioeconomic responses. Where possible, the literature cited in this chapter is specific to changes in the United States. KEY MESSAGES.
Key Message 1: Land-Cover Changes Influence Weather and ClimateChanges in land cover continue to impact local- to global-scale weather and climate by altering the flow of energy, water, and greenhouse gases between the land and the atmosphere (high confidence). Reforestation can foster localized cooling (medium confidence), while in urban areas, continued warming is expected to exacerbate urban heat island effects (high confidence). Description of evidence baseThe Land-Use and Climate, IDentification of robust impacts (LUCID) project, evaluated climate response to LULCC using seven coupled land surface models (LSMs) and global climate models (GCMs) to determine effects that were larger than model variability and consistent across all seven models. Results showed significant discrepancies in the effect of LULCC (principally, the conversion of forest to cropland and grassland at temperate and higher latitudes) on near-surface air temperatures; the discrepancies were mainly attributable to the modeling of turbulent flux (sensible heat the energy required to change temperature and latent heat the energy needed to change the phase of a substance, such as from a liquid to a gas). Land surface models need to be subjected to more rigorous evaluations, and evaluate more than turbulent fluxes and net ecosystem exchange. Rigorous evaluations should extend to the parameterization of albedo, including the effect of canopy density on the albedo of snow-covered land; the seasonal cycle of albedo related to the extent, timing, and persistence of snow; and the benchmarking of the effect of present-day land cover change on albedo.
More recently, there is consistent modeling and empirical evidence that forests tend to be cooler than nearby croplands and grasslands.,The study of the influence of wildland fire on climate is at its advent and lacks a significant knowledge base., Improved understanding would require more research on the detection of fire characteristics; fire emissions; and the relative roles of greenhouse gas (GHG) emissions, aerosol emissions, and surface albedo changes in climate forcing.The urban heat island (UHI) is perhaps the most unambiguous documentation of anthropogenic modification of climate. Two studies have found that the stunning rate of urbanization in China has led to regional warming, which is consistent with the observation that land-use and land-cover changes must be extensive for their effects to be realized. Research on the effects of urbanization on precipitation patterns has not produced consistent results., Uncertainties related to the effect of urban areas on precipitation arise from the interactions among the UHI, increased surface roughness (for example, tall buildings), and increased aerosol concentrations. In general, UHIs produce updrafts that lead to enhanced precipitation either in or downwind of urban areas, whereas urban surface roughness and urban aerosol concentrations can either further contribute to or dampen the updrafts that arise from the UHI. Major uncertaintiesLand use and land cover are dynamic; therefore, climate is influenced by a constantly changing land surface. Key Message 2: Climate Impacts on Land and EcosystemsClimate change affects land use and ecosystems.
Climate change is expected to directly and indirectly impact land use and cover by altering disturbance patterns ( medium confidence), species distributions ( medium confidence), and the suitability of land for specific uses ( low confidence). The composition of the natural and human landscapes, and how society uses the land, affects the ability of the Nation’s ecosystems to provide essential goods and services ( high confidence). Description of evidence baseMuch of the research assessing the impact of climate change on agriculture has been undertaken as part of the Agricultural Model Intercomparison and Improvement Project (AgMIP), which has been understandably focused on productivity and food security., Less effort has been devoted to understanding the impact of climate change on the spatial distribution of agriculture. Deryng et al.
(2011) used one of the AgMIP crop models (PEGASUS) to show poleward and westward shifts in areas devoted to corn, soybean, and wheat production. Parker and Abatzoglou (2016) have reported a poleward migration of the USDA’s cold hardiness zones as a result of a warming climate.
Several empirical studies have found an increase in wildland fires in the western United States over the last several decades, in which indicators of aridity correlate positively with the amount of area burned.
BackgroundLand use/cover (LULC) change is a dynamic and complex process that can be caused by many interacting processes ranging from various natural factors to socioeconomic dynamics. It exerts a strong influence on the structure, functions and dynamics of most landscapes. Monitoring and mapping of LULC dynamics are crucial as changes observed reflect the status of the environment and provide input parameters for optimum natural resources management and utilization. The objective of this study was to quantify the spatio-temporal LULC dynamics using satellite image coupled with local perceptions in the Gedalas watershed of the Blue Nile Basin, North Eastern Ethiopia. Maximum likelihood supervised image classification technique were employed to classify LULC categories.
After ensuring acceptable accuracy value for each classified image, image differencing approach was used to detect and quantify LULC transitions of the area. Classification results were validated with the aid of field work, topographic maps, and high resolution Google earth images supplemented with other available thematic data sets. The resultsThe result demonstrated seven major LULC classes and the overall scenario presented by the study reveals that the watershed has experienced quite visible LULC transitions that seem to be continued in the future due to eternal anthropogenic activities and natural factors.
The study ascertain that though there was change in all land use types, the major change detected was a consistent expansion of farmland/settlements area mainly at the expense of Afro/sub Afro alpine vegetation areas. On the contrary, Afro/sub Afro alpine vegetation showed a consistent net loss of over the study of periods.
The findings also highlighted that transitions were ultimately driven by the interplay of biophysical, socioeconomic and institutional factors. Perceptions of the local communities on the LULC change substantially agree with data from satellite images. This implies that the ongoing rural land administration and natural resource conservation and management strategies could not effectively address the expansion of agricultural land towards fragile and marginal lands in the study area. Land resources, which are an integral component of the watershed ecosystem, are essential natural assets which provide social, economic and ecological functions to sustain livelihood of the inhabitants (Nunes and Auge ).
It is the platform on which human activities take place and also the source of materials needed for these activities (Briassoulis ). However, land is becoming a limited resource subject to competing demands and its various functions and services are seriously compromised by the problem of human induced land degradation (Gessesse et al. ).One of the prime prerequisites for better use and management of land resources is information on existing land use/cover patterns and changes in land use/cover through time (Anderson ).
Spatial and temporal status of the land use/cover of a given area is an important parameter in understanding the interactions of the human activities with the environment (Anil et al.; Etefa et al. Land use and land cover patterns change in keeping with demands for natural resources (Anderson ). Studies have shown that although the evidences of land use/cover changes dates back many 1000 years, the recent rates, extents and intensities of human pressure on land and its scarce resources is more rapid and extensive than in any comparable period of time (Petit and Lambin; MEA; Ellis and Pontius ). This unprecedented human and environment interactions have been verified by LULC changes.Despite LULC change have social and economic benefits; this dynamic and complex process usually has an unintentional interlocked multidimensional consequence upon essential Earth’s ecosystem functions and services at both the small and large scales (Lambin et al.; Turner et al.; Lambin and Meyfroidt ).
Download free nice girls finish fat pdf. Nice Girls Finish Fat Put Yourself First and Change Your Eating. From a therapist and expert in emotional eating, the first book to explore the link between weight. From a therapist and expert in emotional eating, the first book to explore the link between weight gain and women who do too much, complete with proven. In Nice Girls Finish Fat, psychotherapist Karen R. Download Free Nice Girls Finish Fat Pdf Reader. Koenig explains the link between the two and gives. Download Free Nice Girls Finish Fat Pdf To Jpg. Makeup can look intimidating, but it is quite easy to apply once you know what to do. This wikiHow will guide.
For instance, changes in LULC has been shown to have negative impacts on biodiversity (Klenner et al. ), biogeochemical cycling and environmental degradation mainly due to exposure of soil to erosion forces (MEA; Meshesha et al.; Mwehia; Sewnet ), stream water quality (Uriarte et al. ); contribution to local and global climate change (Bringezu et al. ) and forest fragmentations (Ligdi et al.; Rands et al.
) which all have implications on the provisioning capacities of the watersheds. Although LULC change is global phenomenon, its nature and magnitude vary from area to areas. Its consequence is, however, strong particularly in fragile ecosystem and mountainous areas.Periodic LULC change monitoring is, therefore, an essential requirement for the assessment of ecosystem health and investigate factors responsible for triggering the dynamic processes and assess the environmental consequences of such changes (Lambin and Geist; Fichera et al.
Moreover, information on LULC dynamics assists in monitoring environmental changes and developing effective land management and planning strategies at both national and local levels (Ellis and Pontius; Etefa et al. ).LULC change has become the focus of geographical research and discourse in Ethiopia for the last few decades (see Zeleke and Hurni; Garedew et al.; Kindu et al.; Temesgen et al.; Meshesha et al.; Desalegn et al.; Ariti et al.; Yesuf et al.; Demissie et al.; Etefa et al. ), but all of these studies were undertaken far from the relatively little known, sensitive and fragile areas of Beshillo catchments. Hence, accurate and up-to-date spatio-temporal information on LULC dynamics, the driving forces and implications of these changes are urgently needed as an input parameter for planning site-specific sustainable land use and resource management practices in such overlooked place. The objective of this research is, therefore, to detect, quantify and map the LULC dynamics and trends as well as drivers and overall socioeconomic and environmental implications associated with these changes by integrating remote sensing information and the insights of the local communities. Methods and materials.
The study area falls into three agro-climatic zones: Temperate (Woina Dega), Cool (Dega) and Wurch (Afro alpine) zone (MOA ) (Fig. The watershed is characterized by a bimodal pattern of rainfall with a 30 years mean annual rainfall of 930 mm. The highest average monthly rainfall (303 mm) was recorded in July; the lowest (10 mm) in December. The rainfall is variable from year to year, both in terms of intensity and distribution. The coefficient of variation of rainfall was calculated at 14.62% and standardize Rainfall Anomaly index also show variation over the period of observation which implying that there is inter annual rainfall variability in the area. The monthly average minimum temperatures range between 6.4 and 8.6 °C while the maximum stretch between 17.9 and 21 °C. The mean annual temperature of the area is 13.6 °C (Fig.
Soil units found in the area are predominantly volcanic in origin. The major soil types of the watershed area are Leptosols, Cambisols, Regosols and Vertisols (Fig.
Though the overall vegetation cover of the watershed are very poor, still there are various types of vegetation which range from scattered woodlots to Afro alpine types. Aside from the patches of eucalyptus trees planted by farmers and government initiated programs; there are a few dispersed indigenous tree species, shrubs and afro alpine species unique to extreme highland parts of the watershed. Fig. 4Dominant soils types of the watershedAt the time of this study, the watershed was occupied by a total population of approximately above thirty-five thousands (Estimation from TWARDO ), which makes a rural crude population density of close to 147 persons per square kilometer; considerably higher than the regional average of 111 persons/km 2 (CSA ).
Agriculture is the main economic activity and source of livelihood. The farming system is almost entirely rain-fed and involves the integration of crop cultivation and livestock rearing that is carried on at subsistence level. Soil tillage is carried out with traditional ard plough (locally called Maresha) pulled by a pair of oxen. The frequency of tillage ranges from 1 to 3 per cropping season depending on crop types.
The major crops of the area are cereal crops, including sorghum ( Sorghum bicolor), Oats ( Avena Sativa), teff ( Eragrostis tef), maize ( Zea mays), barley ( Hordeum vulgare), senar/engedo ( Avena sativa) and wheat ( Triticum vulgare). Barley and senar/engedo are the dominant crops cultivated in the Wurch area of the watershed and sometimes is rotated with wheat.Other major crops are pulses such as Chickpea ( Cicer arietinum), vetch ( Vicia sativa), Field Pea ( Pisum sativum), and Faba Bean ( Vicia faba). Sorghum ( Hordeum vulgare) and Maiz ( Zea mays) are also produced on the lower reach of the watershed.
Four sets of digital satellite imageries such as Landsat MSS, TM, ETM+, and Landsat 8 of the year 1973, 1986, 2001 and 2017, respectively were used to examine spatio-temporal LULC dynamics of the watershed (Fig. The selection of these satellite images was entirely based on the quality (e.g. Cloud free) and long time series availability, correspondence with years of major regime/policy changes and/or events in the country, purpose of the question to be answered, and suitability of the seasons for collecting socioeconomic field data. The Landsat MSS 1973 image had a resolution of 79 m; while the Landsat TM 1986, Landsat ETM+ 2001 and landsat 8 images have a resolution of 30 m (Table ). All these terrain-corrected and radiometrically calibrated Landsat images were freely accessed and downloaded from the online archive of the United State Geological Survey (USGS) Centre for Earth Resources Observation Systems (EROS) available at ( ).
It would have been better to start with medium resolution images such as Landsat TM images for the study but since there were no such images in the 1970s; a lower resolution Landsat MSS 1973 image was taken as reference year for this study. During the acquisition of satellite image and associated metadata, information pertaining to platform, sensor calibration, projection, coverage, resolution, cloud covers and other relevant information were taken into account.
The effects of solar illumination angle induced by the rugged terrain nature of watershed which may cause different reflectance for the same cover type were considered. TM thematic mapper, ETM + enhanced thematic mapper plus, OLI operational land imager, TIRS thermal infrared sensorThe acquisition dates of images were made as close each other as possible to minimize troubles related to seasonal variations. Moreover, dry and clear sky months (December to January) were preferred to mitigate the effects of atmospheric conditions that blight the quality of optical remote sensing imagery. Since these months are seasons of harvesting, they are relatively ideal to easily identify each land use/cover types.All imageries were projected to Universal Transverse Mercator (UTM) zone 37 N using World Geodetic System (WGS) datum (WGS1984UTMZone37N) to ensure consistency between datasets during analysis. The are displayed in the table below. Supplementary dataTo validate classified images accuracy and further investigate driving forces and associated impacts of LULC changes, supplementary data, such as DEM, topographical maps and socio-economic data were collected from various sources.
The Shuttle Radar Topography Mission (SRTM) DEM (30 m spatial resolution) was obtained from the United State Geological Survey (USGS) ( ). A digital elevation model was used for extracting slope properties, prepare topographic map, delineate and identify sub watershed areas, among others. The high-spatial resolution Google Earth was employed to collect the training and testing samples points for land use/cover classification and validation. The 1993 edition three sheets of (1039 A1, 1139 C3 and 1139C4) 1:50,000 scale and 40 m contour interval digital Topographic map were purchased from Ethiopian mapping agency (EMA) and used for geo-referencing of satellite images, preparation of the base maps and for LULC class verification and interpretation.Socio-economic data were collected from the community through a series of questionnaire, key informant interview, focus group discussions which was facilitated by local development agents with the presence of the author and complemented by onsite field observations. The data type captured from local knowledge/experience were the nature of LULC change over time, perceived drivers of such changes and site-specific management efforts.
Additional information supporting the interpretations, including policy documents, proclamations, implementation guidelines and statistical reports were obtained from national and regional government’s offices and official websites. Data analysis was done using SPSS v 22 and MS Excel 2010 Packages while the Satellite image analysis was performed with the help of ArcGIS 10.3 and ERDAS IMAGINE 2014 Softwares. Data verification and error controlData verification and error control have been given special attention in this study. To identify and interpret the main LULC types, field visit and observations were carried out at different times in the watershed. During the field survey, data on the historical and present LULC types were collected and localized by using GPS (GARMIN-60).
Moreover, these reference areas were documented by photographic images (using digital camera) to strengthen the reliability of the data gathered. The GCPs information for the LULC types of the past were identified from elderly people, author’s prior knowledge and knowledgeable assistants from related government departments, such as agriculture and Natural resource offices of Tenta Woreda.
Digital image processing and land use/cover classificationsAfter selection and acquisition of appropriate satellite imagery in terms of coverage, resolution and accompanying metadata from the sources, an intensive preprocessing, such as re-sampling and layer stacking of the required bands were performed. To ensure image pixels uniformity, Landsat MSS 1973 image was re-sampled to 30 m pixels using the nearest neighborhood method (Johnson ). Then, the study area was extracted from the images using Subset tools in Erdas Imagine 2014. After creating subset images covering the watershed and checking the quality of the image, a classification scheme was developed to derive various LULC classes of the watershed.To classify LULC categories, more than three hundred randomly distributed training sets of data, varying in size from 20 to 60 based on area coverage, were employed to locate training pixels for each land use/cover category and study periods. The training sites for recent images were generated by GPS reading supported by high resolution Google earth images while the training sites for older historical images were assigned with the aid of raw images visual interpretation, topographic maps, old black and white ground photographs of some sites within the study area, supplemented by local elder’s information and author’s prior knowledge. This approach was used by other researchers in their study of LULC change in Ethiopian highlands (Kindu et al., Demissie et al.
The training signatures were then evaluated for class separability between LULC classes. In the evaluation stage, signature editions were done by deleting, merging or renaming until the most satisfactory result achieved.Finally, supervised classification and visual interpretation technique were applied for all the images, following Maximum Likelihood classification algorithm and change detection comparison strategies. Authors prefer to use the Maximum Likelihood algorithm with the assumption that it minimises classification error for classes because this approach incorporates both class covariance matrices and mean vectors as signatures, and merging of constituent spectral class signatures to obtain those for rationalized classes (Jensen and Lulla; Foody; Otukei and Blaschke; Kantakumar and Neelamsetti ).In the process of classifications, some LULC units were misclassified to other classes.
For instance, bare lands were misclassified to the farmland/settlements class. This happens due to the fact that some bare land’s spectral properties appeared similar to that of harvested croplands which bring difficulties in distinguishing them during image acquisition. To improve classification accuracy and reduce misclassifications, recoding techniques were applied to clean misclassified ones using ERDAS software through visual inspections. To simplify and reduce the number of LULC classes, some related LULC classes were merged together into one class. For instance, Croplands and settlements were aggregated as farmlands/Settlements as it was difficult to identify the dispersed rural settlements, where in many cases hamlets are surrounded by farmlands, into separate land use/cover types. Similarly, due to the intermingled nature of shrub and bush lands, it was difficult to distinguish these two categories and hence was labeled as shrub/bush lands to report the LULC study results (Table ). Nonetheless, the authors realized that combining the different LULC classes could hidden differences among them and may exert its impact on the quality of the outputs (Misana et al.
For nomenclature purpose, LULC classification types commonly used for Ethiopia were referred and adopted for ease comparison of results. S/NLand use/cover classes aDescription1Woodlands/plantationsAreas covered with intricate mixture of small trees and bushes. These categories also incorporate Eucalyptus woodlots and/or other remnant woody plantations found mainly in the form of farm boundary, in and around scared sites2Shrub/bush landsAreas covered with woody shrubs, thorny bushes and scattered or patches of various species of small trees, usually found along banks of streams, rugged landscapes and escarpments.
Some of these land cover is used for communal grazing and browsing purpose. Then, the overall accuracy, Kappa coefficient, producer’s accuracy and user’s accuracy were calculated from the error matrix (Foody; Fan et al.; Congalton and Green ). The overall classification accuracy was computed by dividing the number of correct values in the diagonals of the matrix to total number of values taken as a reference point; producer’s accuracy was derived by dividing the number of correct pixels in one class divided by the total number of pixels as derived from reference data; user’s accuracy was calculated by dividing correct classified pixels by the total number of pixels and Kappa coefficient, which measures the agreement between the classification map and the reference data, were calculated as per Bishop and Fienberg ( ) (Eq. (1)where K is Kappa coefficient, r is the number of rows in the matrix, x ii is the number of observations in row i and column i (the diagonal elements), x i+ are the marginal totals of row i, x + i are the marginal totals column i, and N is the total number of observations. Change detection and analysisThe changes in LULC that had occurred in the watershed over the period of study were detected through post-classification comparison approach (Singh; Fan et al.; Chen et al. Post classification comparison (map-to-map comparison) methods were preferred in this study for the number of reasons. Firstly, using this method minimizes the problems associated with multi-temporal images recorded under different atmospheric and environmental conditions and by different sensors, (Singh; Yuan et al.
Secondly, this methods provide the extent and nature of ‘from-to’ change information (Jensen ). Based on this ground, four LULC maps were compiled to display the type of LULC classes and compare between the classified images. Then the whole study period (1973–2017) was classified into four sub-periods (1973 to 1986; 1986 to 2001; 2001 to 2017; and 1973–2017 which includes the entire 43 years of study periods). Then, paired overlay was performed through spatial analysis in GIS in order to detect, compare, and analyze patterns and directions of changes and to quantify persistence, gains, losses, total change, net change, and swapping of LULC occurred during the time period considered in the watershed (Pontius et al.
Conversion matrix was used to distinguish the changes of each category at the expense of others and its general structure follows the format displayed on Table. The rows display the categories of time 1 (initial Time), and the columns display the categories of time 2 (subsequent time). Entries on the diagonal (that is, P jj) indicate the amount of land use/cover category which remained persistence of class j between the time period and used to calculate the gains and the losses of land use/cover classes while the off-diagonal entries show the size of the area that transitioned from category “ i” to a different category “ j” during the time interval (Aldwaik and Pontius ). For ease of reference, the equations and notation used to compute various components are presented as follows.
Time 1Time 2Total time 1LossLULC 1LULC 2LULC 3LULC 4LULC 5LULC 6LULC 7LULC 1P11P12P13P14P15P16P17P1+P1+ − P11LULC 2P21P22P23P24P25P26P27P2+P2+ − P22LULC 3P31P32P33P34P35P36P37P3+P3+ − P33LULC 4P41P42P43P44P45P46P47P4+P4+ − P44LULC 5P51P52P53P54P55P56P57P5+P5+ − P55LULC6P61P62P63P64P65P66P67P6+P6+ − P66LULC 7P71P72P73P74P75P76P77P7+P7+ − P77Total time 2P+ 1P+ 2P+ 3P+ 4P+ 5P+ 6P+ 71GainP+ 1 − P11P+ 2 − P22P+ 3 − P33P+ 4 − P44P+ 5 − P55P+ 6 − P66P+ 7 − p77. (10)If the net change is zero (implying gain is equal to loss), then the swap is twice the loss or gain.The exposure of each land use/cover classes for a change were assessed using the loss to persistence ratio (Lp = loss/persistence) which assesses the vulnerability of a land use/cover classes for a change; gain to persistence ratio (Gp = gain/persistence) which evaluates the gain of a land use/cover in comparison to its time 1 size, net change to persistence ratio (Np = net change/persistence) (Braimoh; Ouedraogo et al.
).Values of Gp and Lp greater than one imply that a given land use/cover class has a higher probability to change to other land use/cover class than to persist in its current condition (Braimoh ). If the value of Np was negative, the land use/cover class would have a higher probability to lose area to other land use/cover classes than to gain from them. Finally, two sorts of data were generated; namely, four land use/cover maps, which illustrates the changes in a spatial context and various tables which exhibit the amount of areas for each LULC categories and a cross-tabulation matrix which demonstrate LULC transition from category to category at different study periods. Moreover, bar graphs and tables were used to display quantitative land use/cover class as a function of topography (i.e. Altitude, and slope categories).
The overall methodological flow chart showing the sequence of data acquisition, image classification and analysis is depicted in Fig. The spatio-temporal quantity of LULC types of each category was analyzed in terms of total area and percentage for each study periods (Table, Fig. The result indicated proportion of each LULC classes varied considerably on different dates considered. In 1973, the dominant LULC types in the watershed were farmlands/settlements accounting for more than 50% of the total area followed by Afro/Sub afro alpine vegetation (19.8%), grass lands (16.3%), and shrub/bush lands (10.8%). Bare lands, Woodlands/plantations, and Water courses/beds accounted the smallest proportion of the watershed. Though the extent varied among land use classes, the order of proportion occupied by the LULC types in the study area remained the same in 1986.
In 2001, farmlands and settlements were still the dominant category (48.4%), followed by Afro/Sub afro alpine vegetation (19%) but the ranks of shrub/bush lands (14.8%), and grass lands (12.8%) were reversed. These orders continued for 2017 LULC distributions. The results showed that farmlands/settlements were/are remained by far the dominant LULC of the watershed for the last 43 years in the watershed. Land use/cover type012017Area (ha)% of totalArea (ha)% of totalArea (ha)% of totalArea (ha)% of totalAfro/sub afro alpine vegetation4599.04351314.66Bare lands3971.602.176132.56Farmlands/settlements12,81553.46Grasslands3702.89304012.68Shrub/bush lands2244.80301012.56Water courses/beds1230.10.42910.38Woodlands/plantations17.883.70Total23,9,9,9,970100.00. The overall classification accuracy report of 97, 89 and 96% were attained for the 1986, 2001 and 2017 classified images, respectively. Since the values falls above the cut point of the standard overall classification accuracy level of 85% (Anderson; Congalton and Green ) with no class less than 70% (Thomlinson et al. ), we can conclude that there is an acceptable agreement between the classified image and the ground reality it represents.
A kappa coefficient result was found to be 0.96, 0.78, and 0.94, for years 1986, 2001 and 2017 respectively (Table ). The results showed a strong agreement for each of the three classified images (Lea and Curtis ). UA user’s accuracies, PA producer’s accuracies Land use/cover changes and its driving forces in Gedalas watershed (1973–2017)The LUC changes were categorized into four periods 1973–1986 (first period), 1986–2001 (second period), 2001–2017 (third period) and 1973–2017 (whole period). Change detection diagnostics showed that in the first study period (1973–1986), total area of farmlands/settlements decreased from 50.2 percent to 42.7 percent irrespective of gains made from other LULC types, but then it showed an overall increasing trend throughout the whole study period with the highest change rate being observed in the period from 2001 to 2017 (Table ). The reduction in the first study period was most likely due to abandonment of farmlands by farmers for reasons of massive rural out-migration and resettlement program of the government to mitigate severe land degradation and historic drought episodes of 1984/85 which had a devastating impact on human population and livestock resources of Wello (North Eastern Ethiopia) in general the study area in particular. Such abandoned farmlands were presumably taken over by grass and regenerated into shrub/bush covers hence the observed increase in this two-land use/cover types in this period.
For example, the famine of Wello caused by drought of 1984/85 is still remains as tragic memory to the world community. The following quote taken from the review of Rahmato is a good proof for the recurrent drought problem of the area:The Wello (where the study area is part of it) highlands are on the leeward side of the main rain- bearing winds and thus receive much less precipitation than the highlands in western Ethiopia on the same latitude. Rainfall is frequently unreliable, and on many occasions the belg rains may fail completely, and the Kiremt rains may be short.
This has been the process by which droughts and famines have been triggered for countless generations (Rahmato:6)The Afro/Sub afro alpine vegetation, which was/is the second dominant cover in its surface area, mostly located on the highest elevation and montane ecoregion and accounted for 19.8, 19.4, 19.0 and 14.7 percents of the watershed in the four consecutive study periods. It should be noted that the Afro/Sub afro alpine vegetation areas exhibited declining tendency in its spatial area coverage in all study periods. However, the greatest reduction occurred between 2001 and 2017 compared to that between 1973 and 2001. This could be ascribed to the continu.