3.3. Land Cover and Land Use Changes (LCLUC) and soil erosion changes (2005 - 2015)
In ten years of study period (2005 - 2015) at the national scale in Pakistan, overall ‘cropland’ area increased from 170,761 km2 (19.36%) in 2005 to 177,474 km2 (20.12%) while ‘forest land’ decreased from 2,000 km2 (0.23%) in 2005 and 1,604 km2(0.18%) in 2015. At the national scale, in the ‘grassland’ class 521 km2 (0.06%), in settlements LCLU class 106 km2 (0.01%) and in ‘wetlands/snow cover’ 1,532 km2 (0. 71%) increased detected between 2005 – 2015 while in ‘other’ LCLU class an overall area lost from 551,813 km2 (62.57%) in 2005 to 543,337 km2(61.61%) in 2015 (Table 5).
Through cross tabular or change matrix method, at the national scale total 29,081 km2 (3.30%) loss while 17,506 km2 (2.00%) gain observed in LCLUC classes between 2005 to 2015. Among the LCLU classes, <2% loss and gain detected in ten years of study period (2005 - 2015). In spatial distribution map of LCLU, loss detected prominently in Sindh, Punjab and Khyber Pakhtunkhwa while major gain observed in Balochistan (Table 6 and Figure 6)
Among all six IPCC defined LCLU classes in the seven administrative units of Pakistan, in the Sindh and Balochistan provinces, the area in the percentage of low annual soil erosion (<1 ton/ha/year) category was observed under the majority of the LCLU classes between 2005 to 2015. In Azad Jammu & Kashmir, Gilgit-Baltistan and Islamabad Capital Territory, a well the distribution of each category of soil erosion against LCLU classes were observed between 2005 to 2015. In the forest, under the high (5-20 ton/ha/year) annual soil erosion class, in the Azad Jammu & Kashmir, Gilgit-Baltistan, Khyber Pakhtunkhwa, and Punjab an increasing trend detected in ten years of study period (2005 - 2015). Among all the administrative units of Pakistan, only in Azad Jammu & Kashmir, under the very high (> 20 ton/ha/year) annual soil erosion class, in all six LCLU classes an increment observed in ten years (2005 - 2015) of the study period. Surprisingly, in the settlement class in 2015, very high (> 20 ton/ha/year) soil erosion was calculated as 10% which was zero in 2005. Even in the Khyber Pakhtunkhwa province and Islamabad Capital Territory, under the high (5-20 ton/ha/year) soil erosion class increase in settlement class observed between 2005 to 2015. In the settlement class, in ten years of the study period (2005 - 2015), no soil erosion occurred in the Gilgit-Baltistan. In the Islamabad Capital Territory, from 2005 to 2015, the other LCLU class completely shifter from medium (1-5 ton/ha/year) to high (5-20 ton/ha/year) annual soil erosion category (Figure 7).
At 5km spatial resolution, bivariate analysis revealed in Pakistan maximum areas covered by ‘no change’ between soil erosion change and LCLUC classes. Majority LCLUC ‘loss’ detected in Balochistan, Punjab and Sindh with no-change in soil erosion change. In the Azad Jammu & Kashmir, Gilgit-Baltistan, Islamabad Capital Territory and Khyber Pakhtunkhwa administrative units, ‘loss’ in LCLUC detected with ‘gain’ in soil erosion change class. Huge area in Khyber Pakhtunkhwa and Gilgit-Baltistan spreader under ‘loss’ with few patches of ‘gain’ in LCLUC and soil erosion change between 2005 - 2015 (Figure 8).
4. Discussion
This study provides the first soil erosion estimations and monitoring from 2005 to 2015 at 1 km spatial resolution at the national scale and seven administrative units of Pakistan. In this study, we adopted a cost-effective and easily replicable methodology, used freely available geospatial datasets (soil, precipitation, LCLU, and elevation) in the widely adopted and proven RUSLE.