Pakistan’s Timberline Forest Ecosystem Dynamics vis-a-vis Changing Climate Across Three Provinces (KPK, AJK and GB) Using GIS/RS Techniques

World's forests have always been under anthropogenic threats leading to instantaneous and sound supervision using satellite-based data collecting capabilities through GIS/RS. Timber line in Pakistan consists of sub-alpine forests and alpine pastures of KPK, AJK and GB. Research findings showed that earth's climate was changing rapidly than previously assumed, demanding the spatiotemporal dynamics analysis using GIS/RS techniques in prevalent climate change scenarios. Two separate methods were used; 1) Visual Interpretation of Supervised Classification Technique for AJK & GB and 2) object Oriented Classification of Supervised Classification Technique for KPK. The primary data for T (max), T (min) & Ppt. (1980-2013) was taken from PMD, Islamabad. Province wise data showed 15 % change in forest cover area for timber line in both KPK (142780 Ha) & AJK (24990 Ha), followed by GB (39267 Ha) as13 % across 2005-2014. The results calculated that the average upward shift in timberline was highest in KPK (285 m), followed by AJK (233 m) and GB (170m) across 2005-2014.The data also showed change in avg. T (max) was 3.4°, -1.8° and 0.6° C, avg. T (min) was -2°, -1.5° and -2° C while change in total precipitation was 88.5, -3.7 and 75.5 mm for AJK, GB and KPK respectively. The Pearson's Correlation Co-efficient chart concluded that climatic factors showed a strong and positive correlation among themselves as well as with change in elevation. However, the correlation among climatic factors and change in forest cover area was weak concluding deforestation to be exclusively an anthropogenic phenomena. Change in elevation showed a weak and negative correlation with change in area while all other correlations were non-significant.


Introduction
Land cover data play an important role in understanding and resolving various issues pertaining to the natural resource management including forests. Forests are an integral natural resource and should be preserved on main concern basis for sustainable ecological development. The practical application of GIS/RS was initiated in 1930 by means of aerial photographs and the launching of the first Earth Resources Technology Satellite (ERTS-1) in 1972 led to the monitoring of natural resources including forests in the world using GIS/RS techniques for land cover area analysis. Forest change detection mapping through satellite images was found to be the 'most influential surveying tool' for conservation organizations, local administration and the non-government organizations (Simoonga et al, 2009). Tree cover are changes monitoring was considered of utmost importance for the forest ecosystem directors (Mitasova et al, 1996). Growing levels of human disturbances have urged terrific heaviness on the forests. Therefore, instantaneous supervision and maintenance of the regional forests using modern tools and techniques was declared imperative (Rabindra et al, 2010) because forest ecosystem management requires precise and rationalized resource statistics. Satellite -based data collecting capabilities have the potential to detect variations accurately and mapping can be done in modified ways. Despite the multiplicity of functions, a number of conclusions can be made about the importance of GIS in the field of forestry. Using remote sensing, particularly Light Detection and Ranging (LIDAR) to derive relevant forest characteristics for scientific forest management through Species Distribution Models were concluded to be the need of the hour (Rechsteiner et al, 2017). Mccall and Minang, (2005) critically evaluated the participatory GIS (PGIS) used in participatory mapping and planning for community-based natural resource management. Timber line in Pakistan consists of Subalpine forests and alpine pastures. Alpine Pastures located (above 3350 m) in the north of the country i.e., in northern Hazara, Malakand Division of KPK and Gilgit, Baltistan & Diamir districts of GB These includes alpine meadows and shrub formation including evergreen Junipers macropoda Boiss., Rhododendron arboretum Sm., Ephedra nebrodensis Tineo., Salix tetrasperma Roxb.and Lonicera spp Thunb. Sub-alpine forests (upto 3350 m) are found in AJK, Upper Dir, Swat, Chitral, Hazara (KPK) and Gilgit (GB). Main species are Abies pindrow [(Royle ex D.Don)Royle] and Betula utilis D .Don., with undergrowth of Viburnum nervosum D. Don. and Salix tetrasperma Roxb., (Bukhari et al, 2012).Research findings showed that earth's climate was changing rapidly than previously assumed and researchers were working actively to compute the climate change and its impacts on natural resources particularly forests all over the world by using various modern tools, techniques and methodologies including GIS/RS. Pakistan was rated amongst countries, highly exposed to the impacts of climatic events, particularly because of its arid conditions, agrarian economic base, resource scarcity, over exploitation and mismanagement of natural resources (LEADS, 2009). On the whole, impact of climate change on the forest networks of Pakistan can be depressing (Siddiqui et al, 1991). The study thus emphasizes, the assessment of the spatio-temporal dynamics (horizontal & vertical distributional shifts across 2005-2014) of timberline in three provinces including KPK, GB and AJK using GIS/RS techniques in climate change trends of Pakistan.

Materials and methods
The study area comprised of timber line of Pakistan in three provinces of KPK, AJK and GB. As

Methodology
Google Earth images of 2005 & 2014 for 7 districts of GB and 3 districts of AJK were processed using Google Earth Pro (4.2) software by applying the supervised classification technique. The images were already Geo-referenced and rectified. Boundary of the districts was procured from the Survey of Pakistan as given above. The timberline areas were digitized using Visual interpretation method of supervised classification technique and the processed files were saved as ".kHz" format. These files were imported into Arc GIS (10.3) software in the ".shp" format for further processing.
Area calculation of the timber line for both 2005 & 2014 images was made using the tool Calculate Geometry (Hac) for finding the horizontal shifts.
Digital Elevation Model (DEM) written above was overlaid in Arc GIS and contours of 100 m interval were generated using the Spatial Analyst Tool. Point of highest elevations were marked in each district for both 2005 & 2014 images to find the vertical shifts.
Separate map for each districts were prepared and then fit in to the provincial boundary for their respective provinces using Update Tool for data mergering in Arc GIS software for further analysis.

Methodology
SPOT-5 images of 2005 & 2014 for 7 districts of KPK were procured from SUPARCO, Pakistan in ".img" files and opened in E-Cognition Developer for Object oriented Supervised Classification technique. The images were already Geo-referenced and rectified. Boundary of the districts was procured from the Survey of Pakistan as given above. The images were opened in E-Cognition Developer and various colors segments (Polygons) were generated using Segmentation Tool (Multiresolution segmentation---->Scale Parameter 30 size). For applying the Supervised Classification technique, Classification tool was used and various color segments were assigned rent classes as 1. Forests 2, non-forest. Command of Apply was used to classify the timberline areas into Forested & Non-Forested classes. These files were then exported into Arc GIS (10.3) software in the ".shp" format for further processing. These shape files were overlaid on original image for verification of the supervised classification into Forest & Non-Forests. https://doi.org /10.37358/Rev. Chim.1949 Rev. Chim., 71 (5) Digital Elevation Model (DEM) was overlaid in Arc GIS and contours of 100 meter interval were generated using the Spatial Analyst tool. Point of highest elevations was marked in each district for both 2005 & 2014 images to find the vertical shifts.
Separate map for each districts were prepared and then fit in to the provincial boundary for their respective provinces using Update Tool for data mergering in Arc GIS software for further analysis.

Results and discussions
Province wise data shows 15 % change in forest cover area for timber line in both KPK (142780 Ha) & AJK (24990 Ha), followed by GB (39267 Ha) as13 % across 2005-2014 (10 years) ( Table 1, Fig 1-6). Results of ANOVA suggested a significant difference between the average change of forest cover area in different provinces (p-value<0.05). Fisher's LSD indicated significant difference between KPK and GB whereas AJK was found as non-significant to these ( Figure 7).
Province wise data shows that the average upward shift in timberline was maximum in KPK (285 m), followed by AJK (233 m) and GB (170m) across 2005-2014 (10 years) ( Table 2, Fig 1-6). Results of ANOVA suggested a non-significant difference between the average elevation change in different provinces (p-value<0.05). Fisher's LSD also indicated a non-significant difference between the each pair of provinces too (Figure 8).
Province wise data showed change in avg. T (max) was 3.4°, -1.8° and 0.6° C, avg. T (min) was -2°, -1.5° and -2° C while change in total precipitation was 88.5, -3.7 and 75.5 mm for AJK, GB and KPK respectively (Table 3, Province wise data showed that the average upward shift in timberline was maximum in KPK (285 m), followed by AJK (233 m) and GB (170m) across 2005-2014 (10 years). The results were in consistent with the studies of Mihai (2007) estimated an increase in timberline elevations, Klinge et al. (2003) and Siddiqui (1994) anticipated that alterations in the physio-chemical nature of the earth's climate were liable to encompass a noteworthy blow to the forests and related ecosystems like alpine and sub-alpine forests causing them to move upward towards more northerly/southerly latitudes and up elevational gradient from their present locations. The results were in contradiction with Mamet et al. (2015) who modeled the spatial distribution of forests and concluded that forest areas were extending additionally to south instead of real forest confines.           Another argument for such high values of projected deforestation can be lack of expertise or carelessness in handling the satellite image and Google earth maps, application of GIS/RS software and interpretation of the data etc. It is also quite possible that due to upward shifting of timber line regeneration areas might be excluded from the forest cover because of more resemblance to shrubs rather than trees. The results were in consistency with Jose et al. (2008) tested the current position of timberline using GIS techniques. In case of prevailing timberline, it was found at lower levels as compared to the point, it can be. It can be said safely that rate of decreasing in forest cover area might not had been 45.9% for the last 33 years as calculated in the study but there was definitely a decrease in extent of forests of timberline.
Climatic factors of T(max), T(min) & Precipitation along with change in forest cover area & rise in elevation were used as predictors for spatio-temporal dynamics of timberline in consistency with Weisberg et al. (2013) considered the biophysical factors (slope steepness, shrub cover etc.) to predict timberline changes, Jobbagy and Jackson, (2000) explored the processes that underlie the connection between environmental dynamics and the timberline, and also generally found the thermal regime as the most influential clarification for the potential position of tree line.
In province wise comparison, increase in Avg. T (max) was 2.2° C, decrease in Avg. T (min) is -5.5° C while Precipitation showed a total increase of 160. The correlation among climatic factors included in the study was strong and positive indicating that the change in one will definitely affect the others. Lal et al. (2011) concluded that temperature plus precipitation changes, melting of the glaciers, degeneration of natural and environmental resources were the main events distressing the people of the entire universe. Siddiqui (1997) anticipated that change in the physical and chemical nature of the earth's climate was possible to encompass dominating impact on the forests and related ecosystems particularly in alpine and sub-alpine zones constituting timberline. The correlation among Ppt-change in elevation was again strong and positive indicating that with the change in elevation, the Precipitation will definitely show variation. With change in Precipitation depending upon change in elevation, change in T(max) & T(min) will be complementary. Korner and Paulsen, (2004) recognized that the site of tree line was principally determined by temperature, in addition to various other modulating agents. They further concluded that natural tree lines were climate-driven eco-tones, Schworer et al. (2014) suggested that mountain vegetation was highly susceptible to changing temperature and expected to shift upward towards the poles, simultaneously recent paleo-ecological studies have discovered that localized variation in the uphill shift of timberlines in the European Alps was most likely caused by a climatic gradient, Landclim being a vegetation model used under diverse climatic and physical scenarios around Swiss Alps projected a rapid uphill shift of mountainous vegetation areas by 500 m by the end of proceeding century, Svendsen and Kunkel, (2009) reported that numerous natural habitats were shifting towards poles as a consequence of changing climate, Schumacher and Bugmann, (2006) concluded that climate change with all its aspects had exerted major effects on mountain forests of European Alps, Sujata et al. (2015) has pointed out the changes in forest types and composition and extinction of animal and plant species as an aftermath of climate change. Svajda et al. (2011) wrote that mountainous ecosystems were specifically susceptible to climate change because all aspects of the resource in question (composition, diversity and production) were directly linked with temperature. Bogaert et al. (2011) studied the position of the tree line in a sub-Arctic region at various temporal and spatial scales https://doi.org /10.37358/Rev. Chim.1949 and projected that tree line will rise was a consequence of climate warming, Schworer et al. (2014) suggested that mountainous vegetation was highly susceptible to temperature with an upward shift in tree line. Richardson et al. (2009) said that temperature-based theories appear to be the most reasonable for describing timber line dynamics because of its impact on variety of physiological processes along with micro site environment and ecological facilitation for booming regeneration, which was a precondition for uphill progression of tree line. It concludes that upward shift of forests of timberline of Pakistan was definitely dependent on climatic factors. There exist a negative but week correlation among change in area-change in elevation, concluding that timberline forest area will reduce with rising elevations and vice versa. The correlation among T (max), T (min) & Ppt.change in forest cover area was found poor. It concluded that change in forest cover area was weakly and negatively related to change in elevation while it was poorly related to climatic factors.

Conclusions
Present study highlighted the importance of monitoring of renewable natural resources like forests with the application of ever increasing use of modern GIS/RS tools & methods in recording vegetation land cover area, extracting extent of various forest types, detecting the changes in forest area cover as well as upward shifting over time in prevailing climate change phenomena and under current socioeconomic situations for scientific sustainable future forest management in Pakistan. These conclusions were supported by the scientist all over the world in their researches. The conclusions strongly recommend the wise and efficient amalgamation of GIS/RS techniques in mapping the extent, detecting the ongoing changes and devising appropriate strategies for scientifically organizing the natural resources. The resultant findings can be used in improved scientific forest management leading to enhanced forest eco-systems conservation (both protection as well as production). Environmental and climatic benefits of forest can also be highlighted purposefully using GIS/RS techniques.
Climate warming was a reality for timberline area of Pakistan as worldwide. Climatic factors including T (max), T (min) & Precipitation were not only connected strongly with each other but also with forest elevational shift. The change in forest cover area has the weakest relation to the variation in climatic factors and seems to be an entirely anthropogenic matter. Even then, timberline forests were diminishing in area and shifting upward was a fact. This reduction in forest cover will definitely result in reduced timber production and concerns regarding the applied forest management systems, opening new research avenues for future projections on forest management, preservation and conservation.