This paper develops a model that could be used to visualize and predict the violation of restrictions in a given forest. The violation of restriction in this case is assumed to be the difference in areal extent between two forest cover scenes with time; termed “deforestation”. It analyses the relationship in forest cover changes overtime in Ganye Forest Reserve and Glide Cross Country Farm in Adamawa state, Nigeria. Cadastral maps of the forest reserve and farmland were used as the base maps, while the satellite images served as the spatio-temporal data. Landsat ETM+ images of 2003, 2008 and 2013 were used to identify, determine and estimate the violation of restrictions. The result shows that the violation of restrictions could be reliably determined for both the forest reserve and farmland and forecast made in order to predict future occurrence. It also revealed a continuous deforestation in the forest reserve, while in the farmland regeneration of forest stock was noticed. This information is very vital for forest management, planning and decision making in a viable land administration domain.
The rights in a given forest are held through the natural processes of customary land rights or the government instituted statutory rights. Land right is an informal or formal claim to ownership of land, supported by tenure or duration [
The restrictions imposed on forest rights have never been so challenged as it is today in the management of forest. These challenges are in three forms; firstly, there is the challenge of determining the areal extent of violation of restrictions, secondly there is the problem of having a model that can continually be used to assess the violation of restrictions and thirdly, there is this desired need to have spatial data or information useful in the current land administration domain model (LADM) for the monitoring of these violations. Land administration systems (LASs) are a vital instrument of sustainable development and good governance. This good governance is provided through the concept of spatial data infrastructure (SDI) [
The study area is made up of a forest reserve (FR) and a farmland (FM) located in the Southern half of Adamawa State in Nigeria between latitudes 7˚19'25"N and 9˚12'08"N and longitudes 11˚52'55"E and 12˚47'37"E (
The loss of forest cover as a result of deforestation is a major concern to the environmentalists the world over, studies by the Food and Agricultural Organization (FAO), International Tropical Timber Organization (ITTO) and the World Preservation Fund (WPF) have shown that the world has lost about 80% of its forest cover in recent time [
Land rights and tenure security have a positive relationship with deforestation and have been expressed by many environmentalists as having a tremendous influence on the management of the forest [
In the South America, where there is competition for land between the indigenous population, migrants and squatters [
In the Asia, diverse levels on tenurial rights were identified in Thailand. While there is recognition of community and individuals’ rights in line with global norm, multiple registrations of the rights of families are allowed under the country’s land code, foreigners are denied any right in land [
In Africa, land ownership and tenure is complex, depending on the customary and colonial masters’ statutory rights [
In order to protect the forest from misuse leading to deforestation, restrictions are imposed on the forest and its resources, the imposition of these restrictions are case or jurisdiction-specific.
In the South America, Brazil has categorized their restrictions into two, the legal reserve restrictions (LR) and the Permanent Preservation Area restrictions (PPA). Forest restriction within the thick Amazon is 80%, while in the savannah areas, 30% restriction is imposed, and 20% for rural property household. Restriction is also imposed on forests with slopes of 100% and with a height of 1.8 m above mean sea level, and on hilltops or mountain ridges comprising its 1/3 [
In Asia, Thailand is identified as having absolute restriction on logging, by giving it a blanket ban. The country’s land code as amended in 1999 imposed ceiling on the extent of land that individuals or group can own, from o.8 hectares for residential to 8 hectares for agricultural lands [
In Africa, the DRC 2002 land code while granting rights to the customary communities imposed various restrictions on the use of the forest. Gathering of fruits is limited to 2 to 3 km range from the camps or residents, 30 km is approved for hunting range, which may be moved 4 to 6 times every year. Fallow for agricultural fields may extend up to 5 km, while concession of up to 500,000 hectares is approved for timber exploitation [
The diverse nature of approved land rights across the globe and the imposition of restrictions of different kinds on these rights emphasize the importance attached to monitoring forest rights in order to minimize the violation of these restrictive measures to protect the forest.
Ground Surveyed maps showing the areal extent of the forest reserve (Ganye Forest Reserve) at a scale of 1/25,000 and the farmland (Glide Cross Country (GCC)) at a scale of 1/5000 form the base maps for this study. The forest reserve map was obtained from the Forestry Department of the Ministry of Environment, while that of the farmland was obtained from the Ministry of Lands and Survey, together with their attributes. These maps were obtained in analogue forms. ETM+ images covering the study area for 2003, 2008 and 2013 were downloaded from the United States geological survey website glovis.usgs.gov containing the information given below (
The first step to digital conversion was the scanning of these maps which made them easily compatible with the ArcGIS software.
Each of these analogue maps was imported into the ArcGIS 10.3 for Georeferencing. Georeferencing was done based on predetermined coordinates of identifiable selected points on the maps using the following projection parameters; WGS 1984, UTM Zone 33, Northern Hemisphere. Enough points were added in order to aid good insitu of position points and digitizing. The areas encompassing the forest reserve and farmland were digitized forming vector layer polygons which had hitherto been in raster format (
The digitized maps were then validated by adding them to the open street map (OSM); an editable online map of the globe initiated through the work of the Volunteered Geographic Information (VGI) services or crowdsourcing [
Preprocessing is done to correct image irregularities spectrally, radio metrically, spatially and geometrically before it can be finally fit for any classifications.
The 2003 and 2013 images, which can be termed, scan line corrector on (SLC-on), which means the scan line corrector was functional and did not reveal any scan gap were preprocessed for their spectral enhancement through composite bands or layer stacking, which enabled viewing of features from different spectral band combinations. This was done using the ERDAS IMAGINE 2014 software. They were spatially enhanced by statistical filtering which enable distinguishing of objects within a given space. Radiometric enhancement was done by applying the histogram equalization which maximizes contrast between pixel values.
The 2008 image is a SLC-Off image, which means the scan line corrector malfunctioned, needed to be spatially
Year | Path/Row | Band Combination | Acquisition Date |
---|---|---|---|
2003 | 185/54 | 4, 3, 2 | 19/01/2003 |
2008 | 185/54 | 4, 3, 2 | 18/02/2008 |
2013 | 185/54 | 5, 4, 3 | 12/04/2013 |
corrected before any other enhancement can be carried out. Although there have been reservations about the scientific applications of SLC-Off images, it has been proven to be useful in forest and vegetation studies [
The final process of processing the satellite imageries were done by classifying the images into various themes. The first step involved the use of pre-selected training sites to obtain the average spectral values or digital numbers (DN) that would be used for the classification. These sites were selected during “ground truthing” or site visitation, a process normally used to validate classification accuracy. In the ERDAS IMAGINE 2014, polygon signatures were created for the classes of water body, green forest, dry forest, built-up areas, mountains/outcrops and bare soil (
Accuracy evaluation was carried on the classified images using the reflectance value of the ground truthed sample sites signatures. The error matrix show insignificant errors which cannot significantly affect the accuracy of the classified themes (
The superimposition of the digitized and validated shape file maps of the study areas on the satellite images in ERDAS IMAGINE 2014 to create their subsets was the first form of data integration.
In order to simplify the excesses of having too much vector polygons for parcel conversion and editing, Adamawa state and the selected study sites within it were carved out of the main raster image scenes as subsets or
Classified Data | Green Forest | Dry Forest | Bare Soil | Built-up A | Water | MT-Outcrop | Row Total |
---|---|---|---|---|---|---|---|
Green Forest | 489 | 0 | 0 | 1 | 8 | 0 | 498 |
Dry Forest | 0 | 77 | 0 | 0 | 0 | 0 | 77 |
Bare Soil | 0 | 0 | 386 | 0 | 0 | 0 | 0 |
Built-up A | 0 | 0 | 0 | 2437 | 16 | 0 | 2453 |
Water | 0 | 0 | 0 | 0 | 1500 | 0 | 1500 |
MT-Outcrop | 0 | 0 | 0 | 0 | 0 | 1573 | 1573 |
Row Total | 489 | 77 | 386 | 2438 | 1524 | 1573 | 6487 |
Classified Data | Green Forest | Dry Forest | Water | Built-up A | Bare Soil | MT-Outcrop | Row Total |
---|---|---|---|---|---|---|---|
Green Forest | 1072 | 1 | 26 | 11 | 1 | 7 | 1118 |
Dry Forest | 2 | 903 | 0 | 0 | 0 | 0 | 905 |
Water | 0 | 0 | 2213 | 0 | 0 | 0 | 2213 |
Built-up A | 0 | 0 | 4 | 3420 | 0 | 0 | 3424 |
Bare Soil | 0 | 0 | 0 | 14 | 1147 | 0 | 1161 |
MT-Outcrop | 0 | 0 | 0 | 0 | 0 | 4416 | 4416 |
Row Total | 1074 | 904 | 2243 | 3445 | 1148 | 4423 | 13,237 |
Classified Data | Green Forest | Dry Forest | Built-up A | Bare Soil | MT-Outcrop | Water | Row Total |
---|---|---|---|---|---|---|---|
Green Forest | 3832 | 0 | 0 | 0 | 0 | 10 | 3842 |
Dry Forest | 1 | 378 | 0 | 0 | 0 | 0 | 379 |
Built-up A | 0 | 0 | 1760 | 0 | 0 | 95 | 1855 |
Bare Soil | 0 | 0 | 0 | 1202 | 0 | 0 | 1202 |
MT-Outcrop | 0 | 0 | 0 | 0 | 1050 | 4 | 1054 |
Water | 0 | 0 | 0 | 0 | 0 | 106,332 | 106,332 |
Row Total | 3833 | 378 | 1760 | 1202 | 1050 | 106,441 | 114,664 |
areas of interest (AOIs) (
The extraction of the themes into vector parcel polygons was done in ArcGIS 10.3 using the attribute table. In the attribute table, a theme is selected based on its “value” (thematic value based on classification) and exported to the data frame to be saved in a choice folder. This process was carried out for each of the image scenes, to produce less complicated vector polygons that could be used for thematic parcel merging, extraction (clipping) and differencing (overlay).
Since the main interest of the study is the forest, the different kinds of forests were merged into a single polygon theme, “forest”. Clipping is done to detect changes in aft and prior polygon features; it shows where changes have occurred and where they have not through area intersections. Differencing of the polygons (forest) was carried to show changes that have occurred between the first and the second image.
The violation of restrictions is determined through the identification and the estimation of areal extent change between successive forest covers. For the three periods of 2003, 2008 and 2013 under study, the images in
The numerical values and percentage of these extracted polygons are presented in Tables 5-7 for the forest reserve.
Summarizing Tables 5-7 for the parcel values of the green and dry forests, in order to get the total forest cover for the three periods give the values in
OID | Value | Count Value | Sum_Area_Sqm | P_Cent | Class_name |
---|---|---|---|---|---|
0 | 3 | 181 | 46,781,899 | 78 | Green Forest |
1 | 4 | 808 | 3,598,658 | 6 | Dry Forest |
2 | 5 | 48 | 109,130 | 0 | Bare Soil |
3 | 6 | 183 | 603,360 | 1 | Built-up Area |
4 | 8 | 246 | 8,619,761 | 14 | Mt_Outcrop |
OID | Value | Count Value | Sum_Area_Sqm | P_Cent | Class_Name |
---|---|---|---|---|---|
0 | 3 | 1325 | 9,183,534 | 15 | Green Forest |
1 | 4 | 353 | 5,686,115 | 10 | Dry Forest |
2 | 6 | 224 | 786,503 | 1 | Built-up Area |
3 | 7 | 36 | 207,342 | 0 | Bare Soil |
4 | 8 | 176 | 43,845,760 | 73 | Mt_Outcrop |
OID | Value | Count Value | Sum_Area_Sqm | P_Cent | Class_Name |
---|---|---|---|---|---|
0 | 3 | 168 | 506,266 | 1 | Green Forest |
1 | 4 | 164 | 493,448 | 1 | Built-up Area |
2 | 6 | 478 | 10,568,146 | 18 | Dry Forest |
3 | 7 | 417 | 2,010,053 | 3 | Bare Soil |
4 | 8 | 263 | 46,132,893 | 77 | Mt_Outcrop |
Class Name | 2003 | 2008 | 2013 |
---|---|---|---|
Green Forest | 46,781,899 m² | 9,183,534 m² | 506,266 m² |
Dry Forest | 3,598,658 m² | 5,686,115 m² | 10,568,146 m² |
Total | 50,380,557 m² | 14,869,649 m² | 11,074,412 m² |
In a similar manner, the numerical values and percentage of these extracted polygons parcels for the farmland are presented in Tables 9-11.
The summary of extracted forest parcels from Tables 9-11 for the farmland is given in
In modelling Restrictions pertaining to deforestation, the “process approach” is used. Chen [
The violation of restrictions between 2003 and 2008 for Ganye forest Reserve is given by the total sum difference of polygon parcels between the two periods (
Similarly, the violation of restrictions for the period between 2008 and 2013 can also be obtained by the difference in the forest cover between these periods (
In a similar manner, the total forest cover (green + dry) for the three periods of 2003, 2008 and 2013 as summarized in
Plotting these values give a clear picture of the trends in the forest cover as shown in
OID | Value | Count Value | Sum_Area_Sqm | P_Cent | Class_Name |
---|---|---|---|---|---|
0 | 3 | 6 | 12,600 | 3 | Green Forest |
1 | 4 | 2 | 362,925 | 83 | Dry Forest |
2 | 5 | 15 | 52,089 | 12 | Bare Soil |
3 | 6 | 8 | 11,476 | 3 | Built-up Area |
OID | Value | Count Value | Sum_Area_Sqm | P_Cent | Class_Name |
---|---|---|---|---|---|
0 | 3 | 40 | 133,543 | 30 | Green Forest |
1 | 4 | 9 | 222,637 | 51 | Dry Forest |
2 | 7 | 6 | 82,463 | 19 | Bare Soil |
OID | Value | Count Value | Sum_Area_Sqm | P_Cent | Class_Name |
---|---|---|---|---|---|
0 | 4 | 1 | 450 | 0 | Built-up Area |
1 | 6 | 5 | 364,725 | 83 | Dry Forest |
2 | 7 | 14 | 73,801 | 17 | Bare Soil |
Class Name | 2003 | 2008 | 2013 |
---|---|---|---|
Green Forest | 12,600 m² | 133,543 m² | |
Dry Forest | 362,925 m² | 222,637 m² | 364,725 m² |
Total | 375,525 m² | 356,280 m² | 364,725 m² |
Forest Cover | Year |
---|---|
50,380,557 m² | 2003 |
14,869,649 m² | 2008 |
11,074,412 m² | 2013 |
The areal extent or amount of violation of restrictions (deforestation) can be estimated by subtracting the value of the later from the former or previous cover. This areal extent, which is actually the deforestation, is shown in red against the preceding year’s (PY) value in blue (
The violation of restrictions in farmland is determined through the same process as shown in the data and processes below (
Applying the same procedure as that of the forest reserve, the values of forest cover for the farmland for the periods of 2003, 2008 and 2013 were rearranged to obtain a summarized forest cover shown in
Plotting these values give the chart in
The areal extent or amount of violation of restrictions (deforestation) in the farmland is shown in
Thus, the amount of violation of restrictions between 2003 and 2008 is put at 19245 m², while between the periods of 2008 and 2013; there was improvement in forest cover of 8445 m². This does not however indicate that there was no deforestation between these periods.
The usefulness and validity of the models are tested by using Pearson linear regression to analyze the directionality and correlation of the data. Regression analysis gives the relationship between two (or more) variables; one called the response variable and the other called the explanatory or predictor variable [
In the case of this research, the temporal data (years) is the independent variable (x) that is used to predict the spatial data or information (forest cover) which is the dependent variable (y). Understanding the relationship between these variables is important in order that future predictions may be made by simple forecasting.
This simple forecasting for the forest reserve predicts a negative value for 2018 (−13,864,605.67 m²) (
Forest Cover | Year |
---|---|
375,525 m² | 2003 |
356,280 m² | 2008 |
364,725 m² | 2013 |
Forest Cover | Year | |
---|---|---|
Forecast | 50,380,557 m² | 2003 |
14,869,649 m² | 2008 | |
11,074,412 m² | 2013 | |
−13864605.67 m² | 2018 |
This means that by 2018 we would reach an emergency situation where the entire forest reserve would have turned to desert. To buttress this argument, the graph of correlation (
In the case of the farmland, the forecast predicts a positive deforestation value of 354,710 m² (
Findings from the research show that the pattern of deforestation or violation of restrictions in the forest reserves
Forest Cover | Year | |
---|---|---|
Forecast | 375,525 m² | 2003 |
356280 m² | 2008 | |
364,725 m² | 2013 | |
354,710 m² | 2018 |
and farmland while having similar pattern of continuous scattered deforestation differ in the following ways;
・ There is a continuous sharp decline in forest cover in the forest reserve owned by the government/community that threatens the entire ecosystem as indicated by the emergency situation occasioned by the negative value (-) for 2018 forecast.
・ In the farmland, owned by individual/group, there is opportunity for improvement in forest stock through regeneration over time as indicated in the forest cover for 2013.
This study highlights on the modelling of violation of restrictions pertaining to deforestation in a given land right. The model focuses on the use of in situ surveyed ground data in conjunction with aerial realities to provide information on the extent of forest cover change over time, useful for determining and estimating the violation of restrictions. The Ganye Forest Reserve and Glide Cross Country Farm in Adamawa state, Nigeria were used as the study sites. The result of the violation of restrictions in the forest reserve and farmland showed similar downward trend in forest cover with time, however, it was noticed that forest regeneration is possible in the farmland, perhaps because of its size and the type of land right (individual ownership). Conclusively, the research shows that it is quite possible to use accurately surveyed data and aerial images to estimate and predict violation of restrictions in any given land right. The central aim is to provide a model that could be able to continually visualize and predict violation of restrictions for decision making in forest management and planning.
Anthony G.Tumba,Shahidah Bte MohdAriff, (2015) Modelling the Violation of Restrictions Pertaining to Deforestation in a Given Land Right. Journal of Environmental Protection,06,1219-1235. doi: 10.4236/jep.2015.611108