Characteristics and drivers of forest cover change in the post-socialist era in Croatia: evidence from a mixed-methods approach

Extensive forests in Croatia represent an important biological and economic resource in Europe. They are characterised by heterogeneity in forest management practices dating back to the socialist planned economy of the pre-1991 era. In this study we investigated the difference in rates of deforestation and reforestation in private- and state-owned forests during the post-socialist period and the causal drivers of change. The selected region of Northern Croatia is characterised by a high percentage of privately owned forests with minimal national monitoring and control. We used a mixed-methods approach which combines remote sensing, statistical modelling and a household-based questionnaire survey to assess the rates of forest cover change and factors influencing those changes. The results show that predominantly privately owned forests in Northern Croatia have recorded a net forest loss of 1.8 % during the 1991–2011 period, while Croatia overall is characterised by a 10 % forest cover increase in predominantly state-owned forests. Main factors influencing forest cover changes in private forests are slope, altitude, education structure, population age and population density. The results also show that the deforestation in private forests is weakening overall, mostly due to the continuation of the de-agrarisation and de-ruralisation processes which began during socialism.     

Understanding deforestation in montane and lowland forests of the Colombian Andes

Colombian Andean forests cover nine million ha. These forests provide an informative case study of mountain deforestation in South America. They are surrounded by tropical lowland forests, and they host most of the country’s human population. This study evaluates the relative importance of human and natural variables in deforestation of the Colombian Andes between 1985 and 2005 using remote sensing methods, geographic information system (GIS) technology and general linear models (GLM). The following factors affected the annual deforestation in the region positively: forced population migration, unsatisfied basic needs, economic activity, crops, pastures, illicit crops, protected areas and slope. Factors having a negative effect were tenure of small land parcels, road density, water scarcity and mean temperature. The results of this study also provide insight into the differences between the dynamics of lowland forests and those of montane forests. Montane forests had a lower annual rate of deforestation than did forests in the lowlands. Socio-economic, demographic and biophysical factors explain overall deforestation rates for the region. However, when altitude variation is taken into account, intraregional differences in the Andes become evident. Deforestation processes differ between those areas adjacent to the high Andean valleys where most of the country’s population concentrates and those areas in the tropical lowlands north, west and east of the Andean chain. Differences between lowland and montane forest dynamics are due partly to the accessibility of forests and differences in wealth and economic activities. In montane forests, deforestation is positively influenced by economic activity, the presence of protected areas and higher slopes. Deforestation in montane forests is negatively affected by tenure of small land parcels, road density, water scarcity and mean temperature. Lowland deforestation rates are more closely related to rural population, pasture percentage, crops, protected areas and temperature. Our results suggest that montane forests appear to be in a more advanced stage of colonisation and economic development, whereas lowland forests are closer to the colonisation frontier and to rapidly growing colonist populations. This study reinforces the idea that although the most common tropical drivers of deforestation are found in the Andes, these drivers operate differently when intraregional differences are considered.     

Land cover changes in the Lachuá region,Guatemala: patterns,proximate causes,and underlying driving forces over the last 50 years

Identifying the patterns of land cover change (LCC) and their main proximate causes and underlying driving forces in tropical rainforests is an urgent task for designing adequate management and conservation policies. The Lachuá region maintains the largest lowland rainforest remnant in Guatemala, but it has been highly deforested and fragmented during the last decades. This is the first paper to describe the patterns of LCC and the associated political and socioeconomic factors in the region over the last 50 years. We estimated spatial and temporal variations in LCC from a random sample of 24 1-km² landscape plots during three time periods (1962–1987, 1987–2006, and 2006–2011) and evaluated how they were related to some important proximate causes and underlying driving forces. During the study period, 55 % of forest cover disappeared, at an annual rate of 1.6 %. The deforestation rate increased from 0.6 % (during the first study period) to 2.8 % (last period), but there was very high spatial variation. Landscape plots located outside conservation areas and close to roads lost between 80 and 100 % of forest cover, whereas the forest cover in landscapes located within protected areas remained intact during the study period. The establishment of new human settlements, roads, and annual crops was the main proximate cause during the first period, but during the second and third periods, open areas were mainly created to establish cattle pastures. Because ~75 % of forest cover has disappeared outside the protected areas, the conservation of this biodiversity hot spot will depend on the expansion of protected areas, and the promotion of forest regrowth and alternative biodiversity-friendly land uses in the landscape matrix.     

Understanding deforestation in the southern Yucatán: insights from a sub-regional, multi-temporal analysis

The southern Yucatán has been identified as a deforestation hot spot. Land-change studies document the extent of forest loss at a regional scale, and case studies provide insights into the drivers of deforestation at the household level. Those studies have paid minimal attention to sub-regional analysis, especially to discrete land-management units above the household level. This analysis of upland forest change addresses the range of variation in deforestation among 96 ejidos (communal lands) and the Calakmul Biosphere Reserve, the two dominant land-tenure and land-management units in the region. Satellite imagery, census, and land-tenure data are used to establish the extent and location of deforestation patterns, and multivariate techniques are employed to identify biophysical and socioeconomic variables that explain such patterns. Results show that, for the 1984–1993 period, deforestation in the southern Yucatán was not as prevalent as implied by its hot spot designation. Three clusters of deforestation are identified. A logistic regression analysis establishes that size of forest holdings, population growth, and location in the precipitation gradient correlate with ejidos that experienced higher deforestation rates than the rest of the land-tenure units. For the 1993–2000 period, conservation programs and changes in the economic context of this “hollow frontier” contributed to reduce deforestation rates and extent. This analysis illustrates the spatio-temporal heterogeneity of much tropical forest change and caution that it should bring to simple formulations of modeling this change and prescribing policies to control it.     

Proximate causes of deforestation in the Bolivian lowlands: an analysis of spatial dynamics

Forests in lowland Bolivia suffer from severe deforestation caused by different types of agents and land use activities. We identify three major proximate causes of deforestation. The largest share of deforestation is attributable to the expansion of mechanized agriculture, followed by cattle ranching and small-scale agriculture. We utilize a spatially explicit multinomial logit model to analyze the determinants of each of these proximate causes of deforestation between 1992 and 2004. We substantiate the quantitative insights with a qualitative analysis of historical processes that have shaped land use patterns in the Bolivian lowlands to date. Our results suggest that the expansion of mechanized agriculture occurs mainly in response to good access to export markets, fertile soil, and intermediate rainfall conditions. Increases in small-scale agriculture are mainly associated with a humid climate, fertile soil, and proximity to local markets. Forest conversion into pastures for cattle ranching occurs mostly irrespective of environmental determinants and can mainly be explained by access to local markets. Land use restrictions, such as protected areas, seem to prevent the expansion of mechanized agriculture but have little impact on the expansion of small-scale agriculture and cattle ranching. The analysis of future deforestation trends reveals possible hotspots of future expansion for each proximate cause and specifically highlights the possible opening of new frontiers for deforestation due to mechanized agriculture. Whereas the quantitative analysis effectively elucidates the spatial patterns of recent agricultural expansion, the interpretation of long-term historic drivers reveals that the timing and quantity of forest conversion are often triggered by political interventions and historical legacies.     

Climate Variability and Deforestation—Reforestation Dynamics in the Philippines

Changes in the conditions in the 'warm pool' in the Pacific region are reflected in the changes in the local climate system of the Philippines. Both El Niñno and La Niñna episodes in the Pacific Oceans introduce high variability into the local climate pattern, especially rainfall, in the Philippines. Whereas El Niñno appears when annual rainfall is ≥10% lower than normal annual rainfall, La Niñna occurs when annual rainfall is at least equal to the normal. About 15.7 million ha of forest cover had been lost between 1903 and 1998, but only 1.64 million had been reforested in the same period, indicating the presence of unbelievably low ecological stability. Apart from this, the denuded forests freed about 8.24 × 10⁹ Mg C into the atmospheric greenhouse pools. Neither deforestation nor reforestation was undertaken with deliberate regard to the occurrences of El Niñno or La Niñna. Very high rates of deforestation were observed to coincide with or precede strong El Niñno or Niñna episodes, thus confounding further the ecological instability of denuded forest systems, especially those with slope ≥18%. Similarly, the reforestation cycle indicates that saplings are at most 5 years old every time an El Niñno or a La Niñna occurs; in most reforestation schemes, saplings are only 1–2 years old when these events occur. These reforested areas are vulnerable to drought in El Niñno years and to high runoff erosion during La Niñna years. Because they are young, saplings in reforested areas dry easily and pose hazards to forest fires, which were observed to destroy larger tracts of forest cover during El Niñno more so when annual rain ≥10% below the normal. In retrospect, the study indicated that had forests been exploited with conscious regard to the recurrence of El Niñno or La Niñna episodes, ecological impacts could at least be toned down. In the same vein, reforestation should have been more successful it were implemented with due considerations to extreme climate variability. Once trees were planted, the weather elements become more crucial than politicians' meddling and other socio-economic factors to the growth and development of reforested sites.     

Indicators of deforestation in the Southern Brazilian Pre-Amazon

In recent years, Amazonian deforestation has become a question of global concern. Deforestation in Amazon is a complex phenomenon in nature and has been related to traditional agriculture expansion. In this work, land use, socioeconomic and conservation indicators, combined with statistical analysis, were used to understand forces associated with patterns of deforestation. This approach was applied in Southern Brazilian Pre-Amazon in Mato Grosso State, which represents an extensive rain forest-savanna ecotone, located in the south border of Amazon biome. Based on data from the last two agricultural censuses (1995/1996–2006), we compared agricultural expansion in this area and Mato Grosso state. Results have shown that 85 % of state deforestation was concentrated in Southern Pre-Amazon and was closely related to increase in number of cattle and pasture area. PCA results pointed that population (92 %), number of cattle (86.5 %), pastures (84.2 %) and tractors (78.4 %) were variables with highest positive correlation to deforestation. It showed that GDP contributes to an individual axis and has a low correlation to deforestation (37.8 %). Conservation units and indigenous reserves also contribute to a single axis and were negatively correlated to temporary crops area. Results revealed a significant reduction in production and commercialization of extractive products in the region, revealing that the main, almost only conservation policy in Mato Grosso remains the creation of Special Areas. We suggest that further studies are necessary to screen development alternatives to simple cutting trees down. It is important to diversify strategies for deforestation control, and development aspects must be more seriously considered to reach a sustainable deforestation control policy.     

Modelling land use change across elevation gradients in district Swat, Pakistan

District Swat is part of the high mountain Hindu-Kush Himalayan region of Pakistan. Documentation and analysis of land use change in this region is challenging due to very disparate accounts of the state of forest resources and limited accessible data. Such analysis is, however, important due to concerns over the degradation of forest land leading to deterioration of the protection of water catchments and exposure of highly erodible soils. Furthermore, the area is identified as hotspot for biodiversity loss. The aim of this paper is to identify geophysical and geographical factors related to land use change and model how these relationships vary across the district. For three selected zones across the elevation gradient of the district, we analyse land use change by studying land use maps for the years 1968, 1990 and 2007. In the high-altitude zone, the forest area decreased by 30.5 %, a third of which was caused by agricultural expansion. In the mid-elevation zone, agriculture expanded by 70.3 % and forests decreased by 49.7 %. In the lower altitudes, agriculture expansion was 129.9 % consuming 31.7 % of the forest area over the forty-year time period. Annual deforestation rates observed were 0.80, 1.28 and 1.86 % in high, mid and low altitudes, respectively. In the high-altitude ecosystems, accessibility (distance to nearest road and city) had no significant role in agriculture expansion; rather land use change appears significantly related to geophysical factors such as slope, aspect and altitude. In the low-elevation zone, accessibility was the factor showing the closest association with agriculture expansion and abandonment. The analysis illustrates that land use change processes vary quite considerably between different altitudinal and vegetation cover zones of the same district and that environmental constraints and stage of economic development provide important contextual information.     

The recovery of the lower montane cloud forest in the Mucujún watershed, Mérida, Venezuela

High deforestation rates in tropical countries continue to reduce forest cover and thereby habitat quantity and quality. However, in some places the forest is recovering and expanding thus offsetting the biodiversity and ecosystem service losses. In order to characterize the forest recovery, land use and land cover (LUC) changes were analyzed using aerial photographs, taken between 1952 and 2009, of a peri-urban watershed in the Andes region of Venezuela. The qualities of the changes were assessed using landscape indices and hemeroby indicators. In that period, the forest cover increased about 18 %, mainly due to abandoned pastures on steep slopes. At the same time, the urban area expanded about 4 % on valley bottoms, while pastures and crop fields were reduced about 20 %. The results also showed that forest patches were aggregating, whereas pastures were fragmenting. A reduction in direct human impacts on forests growing on abandoned pastures resulted in a slight recovery of the lower montane cloud forest structure and plant composition. But non-native species were found in all LUC categories. During the study period, we documented not only forest recovery, but also urban area growth, intensified land use and invasions by non-native species all of which could partially counterbalance the positives of forest recovery.     

Amazonian forest loss and the long reach of China’s influence

This paper explores China’s role in deforestation in Latin America. Brazil’s Amazon region contains vast natural resources including land, timber, minerals and hydroelectric potential. China’s strong economy and large demands relative to domestic supplies of these resources mean that China has become Brazil’s largest trading partner, primarily for natural resources. The paper examines how China influences deforestation in Brazilian Amazonia in a variety of ways, including through the direct influence of Chinese enterprises through land purchases and other mechanisms. This paper finds that the rapid rise in exports of soy and beef products to China are two of the major drivers of Amazonian deforestation in Brazil. The paper further argues that Chinese purchases of agricultural and forest land and Chinese imports of commodities such as timber and aluminum also cause environmental impacts in Amazonia. Chinese financing and investment in Amazonian infrastructure such as railways and mineral processing facilities have additional impacts.     

Combining remote sensing and household level data for regional scale analysis of land cover change in the Brazilian Amazon

Land cover change in the Brazilian Amazon depends on the spatial variability of political, socioeconomic and biophysical factors, as well as on the land use history and its actors. A regional scale analysis was made in Rondônia State to identify possible differences in land cover change connected to spatial policies of land occupation, size and year of establishment of properties, accessibility measures and soil fertility. The analysis was made based on remote sensing data and household level data gathered with a questionnaire. Both types of analyses indicate that the highest level of total deforestation is found inside agrarian projects, especially in those established more than 20 years ago. Even though deforestation rates are similar inside and outside official settlements, inside agrarian projects forest depletion can exceed 50% at the property level within 10–14 years after establishment. The data indicate that both small-scale and medium to large-scale farmers contribute to deforestation processes in Rondônia State encouraged by spatial policies of land occupation, which provide better accessibility to forest fringes where soil fertility and forest resources are important determinants of location choice.     

Effectiveness of protected areas in the Colombian Andes: deforestation, fire and land-use changes

Protected areas (PAs) are one of the most common conservation strategies to halt biodiversity loss. However, little is known about their current effectiveness and how they could behave under projected scenarios of change, especially in the mountain ecosystems of the Andean region. In this study, we assess the role of protected areas against deforestation and fire incidence in the Colombian Andes and how variables associated with biophysical and context drivers such as the size, year of creation and geographic location of the PAs affect their effectiveness. For the long-term, we have addressed the question of how effective these areas will be under three different scenarios of change for the period between 2020 and 2050, including a reference scenario and two scenarios of increasing pastureland and croplands. The results obtained indicated that both deforestation and fire levels were significantly lower inside protected areas than outside their borders. Concerning variables affecting deforestation and fire, only the type of ecosystem affected these processes in this study. The magnitude of forest loss was different between the analysed scenarios and depended on the location of the PAs. The PAs located in the Western Mountain Range presented the largest values of deforestation but the lowest number of fire incidences. This finding could be related to the presence of processes other than fire, such as illegal logging or selective extraction. The largest losses are expected in the PAs associated with the Eastern and Central mountain ranges, whereas the Western Mountain range will have the lowest loss. We conclude that although the Andes Mountains are a highly transformed and dynamic region, the current PAs have been effective barriers for deforestation and fire. However, the creation of new areas and the maintenance of the current PAs should take into account the interaction with the surrounding population, especially in those areas that are more sensitive to change.     

Characteristics of deforestation in the Democratic People’s Republic of Korea (North Korea) between the 1980s and 2000s

There has been a significant lack of land cover change studies in relation to deforestation in the Democratic People’s Republic of Korea (North Korea). The purpose of this study is to characterize deforestation in North Korea through land cover change trajectory and spatial analysis. We used three 30-m gridded land cover data sets for North Korea representing the conditions of the late 1980s, 1990s, and 2000s, respectively, as well as a digital elevation model. We examined the land cover trajectories during the two decades, i.e., which land cover became which at the pixel level. In addition, we calculated topographic characteristics of deforested pixels. Major findings from the study are summarized as follows: (1) net forest loss in North Korea was negligible in the latter decade compared to the former (>5000 km²), whereas other land cover changes were still active; (2) as a result of deforestation, forest land cover became mostly agricultural, particularly in the latter decade (95 %); (3) expansion of agricultural land cover continued during the time, increasing by >42 %; and (4) elevation and slope of deforested areas decreased slightly in the latter decade. The key contribution of the study is that it has demonstrated which land cover became which at the 30-m pixel level, complementing existing studies that examined overall forest stock in North Korea.     

Dependence on forest resources and tropical deforestation in Ghana

In Ghana, forests provide many products on which the local population subsists. However, these resources are depleting due to a variety of factors including agricultural expansion and over-exploitation of forest resources. This paper presents an analysis of the level of local dependence on forest resources and its implications for forest management in Ghana. The paper also outlines the causes of continuing deforestation in the studied region from the perspective of the local residents and discusses what role they could play in addressing the problem. The aim is to share more light on the current causes of deforestation and make suggestions for improved community-based forest management practices that could help to reduce deforestation. Primary data was collected through personal interviews and focus group discussions with 431 household heads randomly selected from three Forest Districts in Ghana. The survey showed that income from agriculture constituted 60% of the average total rural household income. Forest income provided 38% of total household income, and off-farm income 2%. The four most highly ranked causes of deforestation are poverty-driven agriculture, lack of alternative rural wage employment other than farming, household population levels, and conflict in traditional land practices. This shows a shift in the view of local people who in the past were quick to blame logging companies and government policies for deforestation. The majority of the respondents depended on wild animals like snail, bush meat, wild honey and wild and cultivated vegetables. Given the reasons for deforestation, much thought needs to go into agroforestry practices (e.g. snail farming, bee keeping, fish farming, and vegetable production) in efforts to reduce deforestation, which are currently less promoted. Readers should send their comments on this paper to: BhaskarNath@aol.com within 3 months of publication of this issue.     

Biophysical suitability,economic pressure and land-cover change: a global probabilistic approach and insights for REDD+

There has been a concerted effort by the international scientific community to understand the multiple causes and patterns of land-cover change to support sustainable land management. Here, we examined biophysical suitability, and a novel integrated index of “Economic Pressure on Land” (EPL) to explain land cover in the year 2000, and estimated the likelihood of future land-cover change through 2050, including protected area effectiveness. Biophysical suitability and EPL explained almost half of the global pattern of land cover (R ² = 0.45), increasing to almost two-thirds in areas where a long-term equilibrium is likely to have been reached (e.g. R ² = 0.64 in Europe). We identify a high likelihood of future land-cover change in vast areas with relatively lower current and past deforestation (e.g. the Congo Basin). Further, we simulated emissions arising from a “business as usual” and two reducing emissions from deforestation and forest degradation (REDD) scenarios by incorporating data on biomass carbon. As our model incorporates all biome types, it highlights a crucial aspect of the ongoing REDD + debate: if restricted to forests, “cross-biome leakage” would severely reduce REDD + effectiveness for climate change mitigation. If forests were protected from deforestation yet without measures to tackle the drivers of land-cover change, REDD + would only reduce 30 % of total emissions from land-cover change. Fifty-five percent of emissions reductions from forests would be compensated by increased emissions in other biomes. These results suggest that, although REDD + remains a very promising mitigation tool, implementation of complementary measures to reduce land demand is necessary to prevent this leakage.     

Land change in the southern Yucatán: case studies in land change science

The southern Yucatán Peninsular Region project was designed from the outset as an integrative, multidisciplinary program of study examining tropical deforestation in the largest track of seasonal tropical forest remaining in Mexico and in which smallholder agriculture and a major biosphere reserve are juxtaposed in regard to land uses and covers. Treating land as a coupled human–environment system, the project joins the remote sensing, environmental, social, and modeling sciences in a way that is now recognized as land change science. This paper introduces the project, the study region, and six papers that explore some of the coupled system dynamics in the region. These include the sub-regional variation in deforestation, the pan-regional adoption or anticipation of cattle ranching, the emergence of divergent household agricultural and overall livelihood strategies, the roles of cultural and household histories in agricultural livelihood choices, the temporal intensification of swidden cultivation and its implications for forest species, and carbon stocks across cultivation units, including a new econometric modeling application to forecast changes in these stocks.     

Population,Land Use and Deforestation in the Pan Amazon Basin: a Comparison of Brazil,Bolivia, Colombia,Ecuador, Perú and Venezuela

This paper discusses the linkages between population change, land use, and deforestation in the Amazon regions of Brazil, Bolivia, Colombia, Ecuador, Perú, and Venezuela. We begin with a brief discussion of theories of population–environment linkages, and then focus on the case of deforestation in the PanAmazon. The core of the paper reviews available data on deforestation, population growth, migration and land use in order to see how well land cover change reflects demographic and agricultural change. The data indicate that population dynamics and net migration exhibit to deforestation in some states of the basin but not others. We then discuss other explanatory factors for deforestation, and find a close correspondence between land use and deforestation, which suggests that land use is loosely tied to demographic dynamics and mediates the influence of population on deforestation. We also consider national political economic contexts of Amazon change in the six countries, and find contrasting contexts, which also helps to explain the limited demographic-deforestation correspondence. The paper closes by noting general conclusions based on the data, topics in need of further research and recent policy proposals.Readers should send their comments on this paper to: BhaskarNath@aol.com within 3 months of publication of this issue     

Using multitemporal Landsat TM imagery to establish land use pressure induced trends in forest and woodland cover in sections of the Soutpansberg Mountains of Venda region, Limpopo Province, South Africa

Globally, tropical forests are being perturbed by human activity. Tropical vegetation constitutes some of the largest terrestrial carbon stocks against the build up of greenhouse gases. In this paper, a local-scale case study utilising remote sensing methodology in estimating forest loss is presented, for a section of tropical South Africa’s Soutpansberg Mountains where land use pressure threatens some of the last remaining indigenous forests. Landsat TM images from October 1990, August 2000 and September 2006 were used, together with municipality level demographic data. Hybrid image classification techniques distinguished forest cover on the images, which were classified into vegetation density categories. About 20% of forest and woodland cover was lost in the 16-year analysis period, mainly due to pine and eucalyptus plantation and residential housing expansions. The local-scale key drivers behind the deforestation are examined.     

Agricultural adjustment,population dynamics and forests redistribution in a subtropical watershed of NW Argentina

Patterns of land-use and land-cover change are usually grouped into one of two categories defined by the dominant trend: (1) deforestation resulting from expanding agriculture and (2) forest expansion, usually related to the abandonment of marginal lands. At regional scale, however, both processes can occur simultaneously even in the absence of net change. Given the focus on net change, such redistribution of agricultural and natural and seminatural lands has been generally overlooked. The interaction between agriculture modernization, human demography and complex topographic gradients of northwestern Argentina has resulted in processes of both forest recovery and deforestation, thus providing the opportunity to analyze patterns and driving forces of land-cover redistribution. We analyzed 20 years (1986–2006) of land-cover change in a subtropical watershed in relation to topographic and demographic variables. Although net forest change represented <1 %, forests redistribution affected 7 % of forest lands. There was a consistent geographic segregation of deforestation and forest recovery, with forests expanding over steep highlands and agriculture expanding over lowland irrigated areas. Population trends were not associated to forest expansion in lowlands but they explained 32 % of forest recovery in highlands. Highland forest expansion and lowland deforestation, respectively, imply conservation opportunities for humid montane forests and the environmental services they provide (e.g., watershed conservation) and threats for the conservation of dry forests and its biodiversity. Our study exemplifies the importance of land-use redistribution (rather than net change) with relevant environmental consequences at regional scale.     

三峡库区坡耕地利用与水土保持种植制

三峡库区耕地资源十分有限,且坡耕地占有相当大的比重,陡坡地多,在不良的耕作方式下肥力低下,水土流失十分严重在分析三峡库区自然条件与坡耕地资源特点的基础上,讨论了坡耕地种植制度与土壤侵蚀的关系,提出通过合理搭配作物种类进行多熟制间套作,一年生农作物与多年生经饲作物间作及农作物与木本植物等高间作等手段来建立坡耕地水土保持种植制,从而达到减轻土 蚀之目的。