Land Cover Change

Land cover plays a major role in the hydrology of watersheds. As we have learned from one of our previous posts about hydrologic models (https://bit.ly HydroModelInputs), it is important in the characterization of the topography of the area being modeled.

Natural phenomena and human use of landscape impact the changes in land cover. Different land cover types have different capacities of allowing rainfall to flow as surface runoff. For example, vegetation and trees in agricultural lands and forests intercept water that eventually evaporates through evapotranspiration, leading to less surface runoff. More water are also able to infiltrate the soil in these land cover types. Meanwhile in urban areas where most surfaces are paved, most of the rainfall just flows as surface runoff, which eventually leads to increased river discharges. Consequently, changes in land cover would affect the amount of water in the streams and aquifers. Basically, it would affect the water balance within the watershed.

Rampant land cover change may have significant effects on the peak discharges and baseflow of rivers, as well as on the groundwater recharge to aquifers. Urbanization for example, increases the peak discharges which can result to increased risk of flooding in downstream areas during storm events. Another effect is the decrease in the recharge of groundwater which is an alternative water supply during dry periods.

From NAMRIA land cover maps of 2004 and 2010, there was an increase in urban land cover in Cavite River Basin from 11.1% to 27.3% (approximately 146% increase). This rapid urbanization halted from 2010 to 2015 as a 26.9% urban land cover was recorded in the 2015 Land Cover Map. What if urbanization continued to increase excessively? And what if a huge area of the basin was converted to a closed forest instead?

Let’s look at how increased urban land cover will affect the hydrology of Cavite River Basin. Based on the 2015 land cover maps, and assuming a 2% increase in the urban land cover per year, on the estimate there would be an additional 5,147.5 hectares of urban land in the year 2025.

Paved surfaces increase the surface runoff and the peak discharges of rivers. The long-term hydrologic simulation showed that the annual average peak discharge of the basin increased by about 14 cms (2.3% increase) with the increase of urban land cover area. This flow rate poses an increased risk for flash floods. The annual average surface runoff within the basin also increased by 13.6% or about 3.7 cms. Conversely, the water seeping through the ground that eventually recharges the aquifers are also lessened. The scenario simulation also resulted to a 21.8% decrease (about 2.7 cms) in annual average aquifer recharge. Furthermore, the combined baseflow of rivers (groundwater that discharge to the streams) was also reduced by 21.3% (about 2.3 cms).

On the other hand, if hypothetically a huge part of the basin were converted to forests, the opposite trends can be expected. Since forests allow for more groundwater infiltration, basins that are widely forested yield relatively lower discharges and higher groundwater recharge. In 2010, it was reported that the Cavite province had 40% forest cover. Meanwhile, in the 2015 NAMRIA Land Cover Map, the forest cover within the basin was only at 3.2%. Note that this doesn’t automatically translate to a 36.8% loss. Two main reasons that may explain the large discrepancy are 1) Cavite province is not necessarily the same as Cavite River Basin although the two have a large overlap, and 2) the methods of delineating land cover types between the two agencies (Global Forest Watch and NAMRIA) may have been different.

What if the forest land cover was actually 40%? A scenario was simulated where forest cover was increased to around 40% and lumped in the upstream parts of the basin. The long-term hydrologic simulations showed that the annual average peak discharge of the basin would decrease by around 23 cms (3.7% decrease). The annual average surface runoff would also decrease by around 0.2 cms or 7.9%. Meanwhile, water percolating into the aquifers beneath the ground would increase by 0.1 cms or 10% and groundwater that eventually joins the streams would increase by 0.1 cms or 11%. This implies lower risk of flooding during storm events and higher baseflows in the streams.

For the model inputs, the topographical data used were a resampled 10m-resolution IFSAR DEM for the terrain, 2015 Land Cover Map from NAMRIA, and a global FAO Soil Map. These maps are needed in delineating the basin, as well as for characterizing the subbasins (e.g. how the watershed responds to the rainfall).

The weather input were daily weather data mainly from PAGASA and ASTI. Since PAGASA gauges had been installed for a relatively longer time than ASTI gauges, the PAGASA data were used to create a weather generator to fill in any missing data from any gauge used in the model. The simulation period selected was from 1995 to 2019.

To ensure that the hydrologic model is a good representation of the real-world system, calibration was done with the aid of discharge data from DPWH. The Maragondon and Panaysayan monitoring stations in Cavite were used for calibration. The years when these two stations had continuous data were set as the calibration period. Resulting R2 value and Nash-Sutcliffe efficiency (NSE) both reached up to 0.69 (values greater than 0.65 considered to signify good fit) for some reaches.