Embedded Files
Leo Zurita, Beth-Sua Carvajal and Mark Mulligan, King’s College London
Silvia Benitez, Juan Sebastian Lozano, Jorge Leon The Nature Conservancy, Quito, Ecuador
0. Introduction
The Daule River and its basin are of major importance for Ecuador due to the important and populated coastal lowland it flows through. Ecuador’s most populous city, Guayaquil, uses the Daule’s water as the source for drinking water. Moreover, all of the croplands in the basin depend on the basin’s water regulation services, to supply enough water all year round for irrigation. Hence, knowing and understanding the current status of the basin and potential ways of maintaining and regulating the hydrological ecosystem services it provides, are important. Modelling is an appropriate method to achieve this due to the size of the basin, the lack of historic and comprehensive empirical data available at the basin scale and the need for scenario analysis.
1. Models
WaterWorld is a spatially explicit, physically-based, globally-applicable model for baseline and scenario water balance that is particularly well suited to heterogeneous environments with little locally available data (e.g. ungauged basins) and which is delivered through a simple web interface, requiring little local capacity for use. The model is ‘self parameterising’ in the sense that all data required for model application anywhere in the world is provided with the model (148 inputs maps). However, if users have better data than those provided with WaterWorld, it is possible to upload these as Geographical Information Systems (GIS) files and use them. Results can be viewed visually within the web browser or downloaded as GIS maps. The model’s equations and processes are described in more detail in Mulligan and Burke (2005) and Mulligan (2013). WaterWorld is a grid-based water balance, water quality and soil erosion, transport and sedimentation model. Water balance is comprised of wind-driven rainfall plus fog and snowmelt inputs minus actual evapotranspiration calculated from the vegetation cover and type. The model can be applied at 1-hectare and 1-square-km spatial resolution using different datasets for application to local and national scales respectively. WaterWorld has been applied at sites throughout the world for estimating hydrological baselines and its in-built scenario generator has been used to estimate the impacts of changes in climate, land cover and use and land and water management.
RIOS (the Resource Investment Optimization System) is developed by the Natural Capital Project (Tallis et al. 2013). This tool introduces a science-based approach to prioritising watershed investments by identifying where activities are likely to yield the greatest benefits for both people and nature at the lowest cost (Vogl et al 2012). By coupling WaterWorld and RIOS we are able to provide a sound biophysical underpinning to our intervention impact analysis (WaterWorld) and a sound economic underpinning to our investment optimisation (RIOS).
2. Context
Results from a WaterWorld 1ha simulation for the five tiles that make up the Daule basin provides the context shown in Fig 1 and Fig 2. Fig. 1 shows these tiles and represents water balance as positive for most of the basin (local inputs>local outputs) and especially high in the NE of the basin, but a small area of negative water balance (local ET greater than local supply, indicating provision from upstream) in the southern lowlands. Fig 2 shows tree cover to be high in the SW and N of the basin but low in the agricultural centre and south of the basin.
*
Figure 1. Water Balance for the Daule Basin derived from WaterWorld.
*
Figure 2. Current Tree Cover fraction (0-100) derived from Landsat GeoCover used as baseline.
3. Methodology
WaterWorld: Sensitivity to intervention
We used WaterWorld v2.91 to examine the hydrological baseline of the Daule basin. We then ran two scenarios of (1) complete afforestation and (2) complete deforestation, both for the whole basin in order to highlight the areas that are most sensitive to these changes. WaterWorld produced changes from baseline to scenario as percentage differences from the baseline, thus highlighting the relative change in the key ecosystem services of interest: reducing sediment deposition (DEP), reducing human footprint on water quality (HF) or reducing soil erosion (ERO).
RIOS: Spatial optimisation of allocation of intervention
These outputs from WaterWorld were then used in the RIOS portfolio generator to help identify the optimal locations for investments in these interventions on the basis of the most positive impact on ecosystem services (as estimated by the WaterWorld output) for the available investment.
To run the RIOS portfolio generator we first defined a budget of US$6,000,000 for implementing the interventions. The estimated cost of ‘protection’ in the study area is US$100/ha., whilst afforestation costs are around US$1,500/ha.. We assumed that US$1,000,000 would be invested on protection and the remaining US$5,000,000 on afforestation. In consequence, the portfolio resulted in 10,000 hectares of natural forests protected and 3,300 hectares of non-natural areas afforested with slight differences in their allocation depending on whether reducing sediment deposition (DEP), human footprint on water quality (HF) or soil erosion (ERO) were the determinants for the spatial distribution of the intervention.
WaterWorld: Impacts of optimal interventions
The resulting grids of optimal locations for each intervention identified by RIOS were then used as WaterWorld Zones of Interest (ZOIs) to examine the local and catchment-wide biophysical impacts of these site-specific interventions. The resulting intervention areas were distributed only within three of the five tiles that the basin occupies within WaterWorld. We ran each scenario separately within is allocated tile and using the spatial investments optimised for reducing sediment deposition (DEP), improving water quality (reducing human footprint, HF) and reducing soil erosion (ERO). These investments were used to run the appropriate scenarios and produce mapped spatial output.
In addition, a group of Points Of Interest (POIs) were identified within the catchment including the main populated towns, dams, and other sites. The mean output values within a 5 pixel buffer around them (a 5 pixel square window around the POI pixel) was calculated, in order to account for geographical imprecision in the coordinates of the POI and the datasets. From these men values in native units were extracted and entered into a spreadsheet. A maximum change for each site across all three services was calculated as well as change and percentage of change between the baseline and the scenarios.
3.1. Services
The key services of interest are:
Hillslope net soil erosion (ERO)
Erosion of soil and sediment from the hillslopes that is transported downstream due to water runoff. The net hillslope soil erosion value used here is calculated as gross erosion minus deposition. It is expressed in mm/yr.
Soil Deposition (DEP)
The amount of sediment deposited within an area (cell) instead of being transported down the channel. This occurs when the amount of sediment held in transportation is greater than the sediment transport capacity. It is also measured in mm/yr.
Water quality (HF)
Human footprint (HF) is a proxy for water quality within WaterWorld as described by Mulligan (2009), and has a value from 0 to 100% of potential pollution based on the distribution of rainfall to human (polluting) land covers and natural (non-polluting) land covers. HF calculates the percent of water in each pixel that fell as rain on potentially polluting land uses (cropland, pasture, urban, roads, mining, oil and gas) upstream and how its impact flows and it is diluted downstream.
3.2. Scenarios
We implemented a series of interventions using WaterWorld: i) afforestation (AFF), ii) protection (PROT) and iii) business as usual (BAU). These scenarios could be applied in the basin to improve environmental protection and conservation objectives, while BAU is used as a reference for future change in the absence of investment to examine the impacts of the PROT scenario. Each scenario is applied in the three tiles according to the distribution of optimal areas for investment as defined by RIOS. This to maximise the positive effects of: reducing sediment deposition (AFFDEP & PROTDEP), reducing human footprint (AFFHF & PROTHF) and reducing net hillslope erosion (AFFERO & PROTERO).
AFF scenario
For the AFF scenarios, we afforested non-natural areas. See table 1 for the rule applied.
BAU scenario
In the BAU scenario we used the WaterWorld land use change model to convert deforested areas from their current land use to a cropping land use, characterised by a per-pixel land cover of 10% trees, 70% herbaceous and 20% bare. Assignment of LUCC was made on the basis of the current road network but also any known planned roads for the areas. See table 1 for the rule applied. WaterWorld’s LUCC model projects recent rates of land use change forward on the basis of tree cover change from terra-i (www.terra-i.org) and MODIS-VCF 2010-2000 (Hansen et al., 2006; Townsend et al., 2011) averaged over local administrative areas, with assignment of the cells to be deforested being dependent upon accessibility to population centres by road and river as well as and proximity to existing forest-agriculture boundaries.
PROT scenario
In the PROT and Business as usual (BAU) land cover and land use change (LUCC) scenario we used the the WaterWorld LUCC model.
PROT examined the possibility of affording protection to the optimal locations for investments defined from the WW-RIOS portfolio. To achieve this we run a similar deforestation LUCC scenario (named PROT) but this time considering the areas identified by RIOS to be protected and preventing deforestation in those areas (which has the impact of shifting the allocation of deforestation elsewhere). The land use change model is run with the same settings as previously but this time we use WaterWorld’s zone of interest (ZOI) tool to exclude deforestation from the “protected” areas. See table 1 for details of applied rules.
Table 1 Rules for the three intervention scenarios
4. Results
4.1. Hydrological baseline
The Daule River, in Western Ecuador, flows from Santo Domingo de los Tsachilas to the coastal city of Guayaquil (Ecuador’s second city and its industrial capital). The basin includes montane areas and the lowland Pacific plains of Ecuador. According to the SRTM DEM used by WaterWorld the area has elevations from 820 masl in the slopes of Santo Domingo to 0 masl down in the Gulf of Guayaquil and covers some 1.5 million hectares (15,810 km2). The upper Daule feeds the large Velasco Ibarra reservoir (also known as Daule-Peripa) which supplies 2.35 million people of metropolitan Guayaquil. Total annual precipitation within the basin varies from 640 to 3200 mm/yr according to WaterWorld’s wind driven rainfall metric.
a) b)
Figure 3. a) Location of the Daule Basin in Ecuador and b) detail of the SRTM Digital Elevation Model of the basin used within WaterWorld.
WaterWorld’s baseline water balance for the catchment varies from -180 to 2700 mm/yr with a mean of 1100 mm/yr and a gradient from the wettest areas in the NW to the driest areas in the S where the water use is higher than the water available, reaching negative values. This ‘shortage’ is covered by flows in from upstream feeding both surface and groundwater uses (Figure 3).
There are no available spatial data on water quality for the basin so we use WaterWorld’s Human Footprint on Water Quality (HF) index (Mulligan, 2009) which examines the potential pollution based on the distribution of rainfall to human (polluting) land and natural (non-polluting) land covers. The HF index in the Daule is on average 32% with the highest values in the upper basin (areas associated with extensive agriculture) and the lowest values in the small mountain regions in the NE and SW portions of the basin. One of the model’s important inputs for this study is the tree cover. It is measured as the fraction of the area (on a pixel basis) that is covered by trees. It reaches 14% on average for the whole basin. This is due to the historic use and extensive human intervention over the land. There are no nationally or internationally recognised protected areas within the basin. There are only a few forest remnants in the upper basin around the reservoir, and in the mountain area in the SW (Figure 3). These areas correspond to the “coastal evergreen lowland forest” group, according to the last ecosystem classification of MAE (2013).
4.2. Scenarios
All scenarios take place within small areas of intervention when compared to the basin total area (1.5 million hectares), and they are interventions on the current land cover (tree, herbaceous, bare soil and land use - croplands). Thus the basin scale impacts are rather low (table 2). However, it is important to recognise the greater positive impacts of each intervention for each service locally, as well as the co-benefits for other services of each intervention. The only intervention with catchment-wide negative impacts in all services is business as usual.
Table 2. Summary of average percentage change for every intervention scenario, highlighting the intended positive changes, and the most negative changes (expected in BAU)
AFFORESTATION
The intervention carried out under this scenario is the afforestation of 3,300 ha. This area amounts to only 0.0002 % of the basin. When aiming to maximise the effects of afforestation by diminishing soil erosion (AFFERO) the mean change for the whole basin was -0.0148%, which is a decrease in erosion that is locally significant in the areas where the forest cover is added and also occurs -to a lesser extent- downstream of these areas. In the AFFDEP scenario, the change in soil deposition across the basin is a decrease on of -0.0052%. This reduction is visible at the local level in afforested areas but is significantly reduced downstream (maps of these areas can be seen in Table 3, second column). The values of human footprint (HF) decrease in all three approaches to the AFF scenario, since land use is converted to “natural” land, thus it is a source of good quality (i.e clean) water compared with the prior land use. It is important to note (table 3, fourth column) that improved quality can be noticed even in the main channel of the Daule, even though the areas of positive intervention are relatively small.
PROTECTION
A total of 10,000 Ha are protected under these scenarios. As explained above and to better illustrate the effects of this protection, the scenario (PROT) is compared with a business-as-usual (BAU) scenario of continued deforestation and agriculturalisation. The PROT scenario runs the same rules, but excludes the areas that are defined as protected from deforestation (thereby increasing allocation outside these areas).
Soil deposition decreases relative to the baseline in PROTDEP by -0.280% over the basin (because of higher water yields under deforestation), whilst the BAU scenario causes an increase in deposition of +0.0207% relative to the baseline and on average for the basin, though directly increasing over an area of only 73,600 hectares. where forest loss leads to greater erosion and sediment transport in excess of transport capacity. Erosion changes in the PROTERO scenario show an increase of +0.0259% which is less than the +0.0322% under BAU. This is a relatively small and local decrease that can be observed only in 753 hectares around the protected intervention areas.
In terms of water quality, under BAU the deforestation and transformation into croplands, leads to the human footprint increasing by +4.53% on average over the basin. However, when protecting the important forest areas (under PROTHF), it increases less, at +4.33%. In general, all these protection scenarios (maps on table 3, fourth column) show, on one hand, negative effects of the land use change model (BAU), but on the other, this is lessened by the positive effects of the protection efforts (PROT).
Table 3. Results of change in services under the scenarios for the Daule Basin for the tiles where areas of intervention (AFF or PROT) were identified by RIOS and modelled in WaterWorld
In order to better summarise the results and show the most effective interventions for a water fund, each scenario should be understood for the basin-scale impact it generates relative to the funds invested. Table 4 summarises these results and shows the area in hectares that is impacted, either negatively or positively, for every million US dollars of investment. The cumulative results are then ranked from 1 to 7 with 1 having the highest areas impacted per dollar and 7 the lowest. The highest positive impacted area per dollar is for the PROT scenarios (ERO>DEP>HF). These also have the highest negative impact (DEP>ERO>HF) per dollar since they do not arrest BAU deforestation outside of the prot areas (the AFF scenario assumes no BAU deforestation)
Table 4 Business case for Daule scenarios
4.3. Points of Interest (POI’s)
The prior analysis takes the whole basin as a unit, which is a comprehensive approach to better understand the impacts of an intervention, even if a that scale, the interventions are relatively small. The analysis of the local changes at the identified points of interest help to pinpoint management relevant impacts at these critical points. Table 5 shows a summary of all the points with the services values for the baseline and scenarios, as well as the percentage of change when compared to the baseline, aiming to highlight the greatest changes, both positive (decreases in ERO, DEP or HF) or negative (increases in DEP, ERO or HF). Furthermore, an interactive map (Figure 4) allows the location and exploration of these points and their main changes.
Table 5. List of points of interest (POIs) showing the changes (in percentage) of each service of interest under every scenario (link to the complete spreadsheet)
Figure 4. Interactive map of the POIs (click over map or here)
4.4. Conclusions
At the catchment scale all scenarios lead to both benefits and to disbenefits in the three ecosystem services of interest, though these benefits and disbenefits occur over very different sized areas and thus are differentially significant (Table 4). The scenario that leads to the greatest total area experiencing service disbenefits is, of course, BAU followed by the PROT scenarios and then the AFF scenarios (in which there is no BAU deforestation). Under the PROT scenarios the area of disbenefits is reduced compared with BAU but disbenefits still occur because deforestation continues outside the PROT areas. The AFF scenario shows very low areas of disbenefits and higher areas of benefits per $ of intervention because in this scenario we focus on afforestation only, not afforestation in the context of ongoing deforestation. PROTERO leads to the greatest area experiencing benefits across all services and this is thus an optimal scenario, followed by PROTDEP and then PROTHF). The AFF scenarios produce the least areas benefitting per $ because whilst HF and DEPOS improve significantly, soil erosion is little affected since a trees provide little more protections against erosion than herb covers.
PROTECTION
AFFORESTATION
BUSINESS AS USUAL
Under BAU deforestation continues removing another 12% of the existing forest cover over the next 50 years. This leads to increases in erosion, sediment deposition and human footprint at the catchment scale. Locally, this leads to increases in soil erosion of >10% for Puerto Limon and increases in human footprint of >10% for Noboa and Bellavista.
The afforestation scenarios affects between 1 and 2% of the basin area and thus their basin-scale impact is low. At points, the AFFERO scenario (protecting areas sensitive to erosion) has little effect on erosion at the points of interest, most of which have little or no erosion under BAU. AFFERO leads to small decreases in human footprint for example at Barraganete. AFFHF leads to decreases in erosion at Puerto Limon >10% and decreases in human footprint at Puerto Limon and Barraganete but these are not >10%. The AFFDEP scenario has little impact on sedimentation at the points of interest but does reduce erosion >10% at Luz America and human footprint is reduced at a number of sites (though by <10%).
The protection scenarios affect between 11 and 13% of the basin area (over a period during which BAU deforestation continues) and thus their basin-scale impact remains relatively low. At points, the PROTERO scenario (protecting areas sensitive to erosion) has no impact (compared with the baseline) on erosion at Puerto Limon, but it leads to increases in human footprint >10% at Barraganete, Chaune, Luz de America and Puerto Limon. This is likely because of continued agriculturalisation outside of the protected areas. PROTHF leads to similar outcomes (increases in human footprint >10% at Barraganete, Chaune, Luz America and Puerto Limon) as well as increases in human footprint >10% at Noboa and Bellavista as per the BAU scenario. Under PROTDEP, there is an minor but important decrease of soil and sediment deposition in Barraganete of >10%, which can be of interest for the Daule-Peripa dam since it is the main point of water intake before the dam. The rest of the changes are similar to those discussed with HF increasing by more than 10% in several POIs, due to further agriculturalisation outside the protected areas.
OVERALL CONCLUSION
(a) Identifying areas that create the greatest change in services of interest and then developing optimal scenarios of land management focused on these areas will have the maximum catchment-wide impact but may not have the optimum impact at particular points of interest downstream. A different approach may be necessary to achieve optimum impacts at points given the complexities of lateral flows in catchments.
(b) Even significant funds like those invested here can only lead to relatively small changes in land use compared with the much larger “business as usual” changes (agriculturalisation) that represent much larger investments. Small changes will have little impact at the catchment scale so they have to be situated carefully relative to specific downstream beneficiaries
(c) Water funds need to be scaled up to match the kinds of changes experienced under BAU and investments in better management of land use (for example reduced impact agriculture) with benefits for ecosystem services and biodiversity may have greater catchment level impacts than investments in small areas of agricultural exclusion, afforestation and protection.
REFERENCES
Hansen, M., R. DeFries, J.R. Townshend, M. Carroll, C. Dimiceli, and R. Sohlberg (2006), Vegetation Continuous Fields MOD44B, 2001 Percent Tree Cover, Collection 4, University of Maryland, College Park, Maryland, 2001
MAE (2013) Ministerio de Ambiente del Ecuador. Mapa de Ecosistemas del Ecuador Continental. Accessed on 28-oct-2013 in http://cdredd.org/sites/default/files/users/common/2013_jointmeeting_ecuador/5MAPA_VEGETACI%C3%93N_Ecuador_19_03_2013.pd
Mulligan, M. (2009) The human water quality footprint : agricultural, industrial, and urban impacts on the quality of available water globally and in the Andean region Proceedings of the International Conference on Integrated Water Resource Management and Climate Change, Cali, Colombia. 11 pp.
Mulligan, M. and Burke, S.M. (2005) FIESTA: Fog Interception for the Enhancement of Streamflow in Tropical Areas. Report to UK DfID, 174 pp. http://www.ambiotek.com/fiesta
Mulligan, M. (2013) WaterWorld: a self-parameterising, physically-based model for application in data-poor but problem-rich environments globally Hydrology Research doi:10.2166/nh.2012.217
Tallis, H.T., Ricketts, T., Guerry, A.D., Wood, S.A., Sharp, R., Nelson, E., Ennaanay, D., Wolny, S., Olwero, N., Vigerstol, K., Pennington, D., Mendoza, G., Aukema, J., Foster, J., Forrest, J., Cameron, D., Arkema, K., Lonsdorf, E., Kennedy, C., Verutes, G., Kim, C.K., Guannel, G., Papenfus, M., Toft, J., Marsik, M., Bernhardt, J., and Griffin, R., Glowinski, K., Chaumont, N., Perelman, A., Lacayo, M. 2013. InVEST 2.5.5 User’s Guide. The Natural Capital Project, Stanford
Vogl, A.L., Heather Tallis, James Douglass, Rich Sharp, Fernando Veiga, Silvia Benitez, Jorge León, Eddie Game, Paulo Petry, João Guimerães, Juan Sebastián Lozano. 2013. Resource Investment Optimization System (RIOS). Software and User’s Guide, v1.0.0. [http://www.naturalcapitalproject.org/RIOS.html]
Page updated
Google Sites
Report abuse