Forecasting hotspots of forest clearing in Kakamega Forest, Western Kenya

Abstract Kakamega Forest is the last remaining rainforest in Kenya and the easternmost remnant of the Guinean-Congolian rainforest belt. As such, it is home to a large number of endemic fauna and flora species. Yet the remaining natural forest is under imminent threat of degradation due to a rapidly growing population in its vicinity and a poverty rate far above the national average. The growing demand for forest resources and ecosystem services will continue to exert great pressure on the remaining forest fragments. In this paper, we predict future hotspots of forest clearing of the remaining natural and old-growth secondary forest in Kakamega Forest. We parameterized an artificial neural network model using resilient backpropagation to simulate the likelihood of forest clearing for each location. Input variables into the network included historic information on forest clearings together with variables capturing the status of forest protection, accessibility to roads and markets, as well as topography and forest density. Simulation results were used to predict future clearings based on observed rates of change. Hotspots of forest clearing were derived by assessing the neighborhood density of predicted clearings. Our results indicate that forest clearings occurred across all forest fragments. Hotspots of future forest clearing tended to occur near roads and market centers. Most future hotspots were found in areas with a lower protection status, where some forest use is allowed. But our model also predicted considerable pressure on remaining old-growth forest resources in the strictly protected National Reserve. Our predictions of deforestation hotspots contribute to a better geographic targeting of nature protection activities and forest management investments in Kakamega Forest. Thus, it will hopefully help policy makers and land managers to strike a balance between satisfying the needs of local livelihoods and preserving the unique ecological values of Kakamega Forest for future generations.

[1]  Pushpam Kumar.  Ecological Sustainability and Human Well Being , 2006 .

[2]  S. Saura.  Effects of minimum mapping unit on land cover data spatial configuration and composition , 2002 .

[3]  S. Phinn,et al.  Characterizing a tropical deforestation wave: a dynamic spatial analysis of a deforestation hotspot in the Colombian Amazon , 2006 .

[4]  J. Kokwaro.  Conservation status of the Kakamega forest in Kenya , 1988 .

[5]  B. Bleher,et al.  Assessment of Threat Status and Management Effectiveness in Kakamega Forest, Kenya , 2006, Biodiversity & Conservation.

[6]  Alan F. Murray,et al.  IEEE International Conference on Neural Networks , 1997 .

[7]  Ellen Shaw,et al.  Roads, Development, and Conservation in the Congo Basin , 2000, Conservation biology : the journal of the Society for Conservation Biology.

[8]  J. Mburu,et al.  Factors Influencing Local Communities’ Satisfaction Levels with Different Forest Management Approaches of Kakamega Forest, Kenya , 2008, Environmental management.

[9]  Iucn Forest Conservation Programme,et al.  Kenya's indigenous forests : status, management, and conservation , 1995 .

[10]  Bryan C. Pijanowski,et al.  Calibrating a neural network‐based urban change model for two metropolitan areas of the Upper Midwest of the United States , 2005, Int. J. Geogr. Inf. Sci..

[11]  E. Lambin,et al.  Proximate Causes and Underlying Driving Forces of Tropical Deforestation , 2002 .

[12]  Eric Koomen,et al.  Comparing the input, output, and validation maps for several models of land change , 2008 .

[13]  J. Barlow,et al.  Quantifying the biodiversity value of tropical primary, secondary, and plantation forests , 2007, Proceedings of the National Academy of Sciences.

[14]  Geoffrey E. Hinton,et al.  Learning representations by back-propagating errors , 1986, Nature.

[15]  M. Seddon,et al.  Land-snail faunas in indigenous rainforest and commercial forestry plantations in Kakamega Forest, western Kenya , 2001, Biodiversity & Conservation.

[16]  Ane Alencar,et al.  Simulating the response of land‐cover changes to road paving and governance along a major Amazon highway: the Santarém–Cuiabá corridor , 2004 .

[17]  A. Angelsen,et al.  Rethinking the causes of deforestation: lessons from economic models. , 1999, The World Bank research observer.

[18]  Helmut Geist,et al.  Our earth's changing land : an encyclopedia of land-use and land-cover change , 2006 .

[19]  W. Leonard,et al.  Local financial benefits of rain forests. Comparative evidence from Amerindian societies in Bolivia and Honduras , 2002 .

[20]  Martin A. Riedmiller,et al.  A direct adaptive method for faster backpropagation learning: the RPROP algorithm , 1993, IEEE International Conference on Neural Networks.

[21]  T. M. Lillesand,et al.  Remote Sensing and Image Interpretation , 1980 .

[22]  C. Metz.  Basic principles of ROC analysis. , 1978, Seminars in nuclear medicine.

[23]  Gertrud Schaab,et al.  Assessing fragmentation and disturbance of west Kenyan rainforests by means of remotely sensed time series data and landscape metrics , 2006 .

[24]  M. Cords,et al.  Long-term tree population dynamics and their implications for the conservation of the Kakamega Forest, Kenya , 2004, Biodiversity & Conservation.

[25]  Eric F. Lambin,et al.  The identification of tropical deforestation fronts at broad spatial scales , 1997 .

[26]  F. Achard,et al.  Pan-tropical monitoring of deforestation , 2007 .

[27]  C. Justice,et al.  MAPPING TROPICAL DEFORESTATION IN CENTRAL AFRICA , 2005, Environmental monitoring and assessment.

[28]  Peter H. Verburg,et al.  Statistical methods for analysing the spatial dimension of changes in land use and farming systems , 2005 .

[29]  Millenium Ecosystem Assessment.  Ecosystems and human well-being: synthesis , 2005 .

[30]  B. Pijanowski,et al.  Using neural networks and GIS to forecast land use changes: a Land Transformation Model , 2002 .

[31]  A. Myrick Freeman,et al.  The Economic Value of Biodiversity , 1998 .

[32]  J. Glenday,et al.  Carbon storage and emissions offset potential in an East African tropical rainforest , 2006 .

[33]  R. Engelman,et al.  Human population in the biodiversity hotspots , 2000, Nature.

[34]  United Kingdom,et al.  GLOBAL FOREST RESOURCES ASSESSMENT 2005 , 2005 .

[35]  R. DeFries,et al.  Detecting Long-term Global Forest Change Using Continuous Fields of Tree-Cover Maps from 8-km Advanced Very High Resolution Radiometer (AVHRR) Data for the Years 1982–99 , 2004, Ecosystems.

[36]  Maureen L. Cropper,et al.  Predicting the Location of Deforestation: The Role of Roads and Protected Areas in North Thailand , 2001, Land Economics.

[37]  Stuart L. Pimm,et al.  Refining Biodiversity Conservation Priorities , 2005 .

[38]  B. C. Pijanowski,et al.  Modelling urbanization patterns in two diverse regions of the world , 2006 .

[39]  David Cieslikowski,et al.  World development indicators 2008 , 2009 .

[40]  R. Marrs,et al.  Conservation in areas of high population density in sub-Saharan Africa , 2007 .

[41]  Thomas M. Brooks,et al.  Time Lag between Deforestation and Bird Extinction in Tropical Forest Fragments , 1999 .

[42]  Jean-François Mas,et al.  Modelling deforestation using GIS and artificial neural networks , 2004, Environ. Model. Softw..

[43]  Heekuck Oh,et al.  Neural Networks for Pattern Recognition , 1993, Adv. Comput..