Documentation

This section shows a brief description of the methodology and information sources used to perform the hydrological simulation in Costa Rica.

For more information on the methodology used and the hydrological model used, the publication by Arciniega-Esparza et al. (2022).

The hydrological simulation database for Costa Rica is available at Zenodo.

Databases

The SRTM (Bamler, 1999) public domain digital elevation model was used with a pixel size of 30m corrected with the national water network as the basis for delimiting the country's catchments. The vegetation cover raster (300m) was obtained from the CCI Land Cover product (Bontemps et al., 2013) and the physical characteristics of the soils were obtained from SoilGrids at 250m (Hengl et al., 2017).

Due to the lack of observed hydrological records, data from different free sources were used to run and evaluate the model. Climate data input to the model comes from CHIRPSv2 (Funk et al., 2015) for daily precipitation and product air temperature from NOAA's CPC Global Daily Temperature product.

The precipitation product was corrected with observed data from 75 available stations in the country using a linear bias correction method (bias = Xobs / Xsatelite). The temperature data were corrected with the elevation of the terrain due to the lack of records for the correction of bias. Global potential evapotranspiration (PET) and current evapotranspiration (ETA) products from the MODIS 16A3 satellite (Mu et al., 2011) were used to adjust the hydrological balance.

The description of the sources used is shown in the following table:

Dataset	Variable	Label	Resolution	Period	Scale	Data type
CHIRPSv2	Precipitation	P	0.05°	1981-preset	Daily	Remote sensing and calibrated merged with ground data
MODIS16	Evapotranspiration	AET, PET	5km	2000-2014	Monthly	Remote sensed
CPC Global Temperature	Temperature	Tmin,Tmax,Tmed	0.5°	1979-present	Daily	Gridded from stations
CCI Land Cover	Land Cover	Land use	300m	1993-2015	Annual	Remote sensed
SoilGrids	Sand and clay content	Soil type	250m	-	-	MachineLearning from soil profiles
SRTM	Terrain elevation	DEM	30m	-	-	SAR interferometry

Hydrological Model

The semi-distributed, conceptual model, based on hydrological processes HYPE (Lindström et al., 2010) was used for the hydrological simulations.

Scheme of the discrimination in sub-catchments of the model and the processes simulated by HYPE. Source: Lindström et al. (2010)

HYPE source code can be downloaded from sourceforge. The official documentation must be used for more information about the model.

Model calibration and performance

HYPE calibration and validation were performed using observed daily flow data in 13 watersheds of Costa Rica, with a calibration period from 1991 to 1999 and a validation period from 2000 to 2003. Additionally, global potential evapotranspiration products were used (PET) and MODIS current evapotranspiration (AET) to adjust the hydrological balance for the period 2001 to 2010, with a validation period from 2011 to 2014.

A step-wise calibration procedure was incorporated for the model calibration, starting with the calibration of the monthly potential evapotranspiration (PET), the monthly evapotranspiration (ETA), the monthly streamflow (Qt) and finally using the daily streamflow (Qt) data. The following figure shows the performance of the model to simulate the monthly streamflow in the 13 monitored catchments.

The simulations correspond to the first version of the regional model of Costa Rica with the HYPE model (HYPE CR 1.0), which was selected from a set of models that generated similar performance with respect to the observed streamflow, however, HYPE CR 1.0 has generated a balanced performance with streamflow (Qt), actual evapotranspiration (AET) and potential evapotranspiration (PET). Likewise, it must be taken into account that there is uncertainty with the HYPE CR 1.0 model since different values of the parameters generated similar results, as can be seen in the following image.

Hydrological Series Simulated

The results of the hydrological model were generated for 605 basins in Costa Rica and cover the period from 1985 to 2019, at monthly and annual scales. The time series of the simulated hydrological variables, as well as the type of statistical aggregation, are shown in the following table:

Variable	Label	Units	Aggregation	Description
Precipitation	P	mm	sum	Bias corrected precipitation from CHIRPS
Average daily streamflow	Qt	m3/s	mean	Total streamflow from the upstream area
Actual Evapotranspiration	AET	mm	sum	Evapotranspiration of vegetation + evaporation of soil and water
Potential Evapotranspiration	PET	mm	sum	Potential Evapotranspiration
Runoff	Qd	mm	sum	Surface runoff (effective precipitation) at sub-catchment scale
Baseflow	Qb	mm	sum	Contribution of water from the three layers of soil to the river at sub-catchment scale
Total flow	Q	mm	sum	Sum of runoff and baseflow at sub-catchment scale
Infiltration	Infil	mm	sum	Infiltration at the upper soil layers
Soil Moisture	SM	mm	mean	Sum of the soil moisture in the three soil layers

Hydrological Indices

Hydrological indices are particularly useful for the synthesis of the long-term hydrological behavior of catchments, allowing the evaluation of spatial patterns in large regions. From the results obtained with HYPE CR 1.0, different hydrological indices were estimated, and are available in the web application.

Name	Label	Description
Evaporative Index	EI	Relates the amount of water that is lost by current evapotranspiration with the water available as precipitation (EI = AET / P)
Aridity Index	AI	Relates the energy available to generate evapotranspiration with the water available as precipitation (AI = PET / P)
Baseflow Index	BFI	Relates the amount of water that the soil layers and aquifers contribute to the hydrological behavior of rivers with respect to the total volume of water in rivers (BFI = Qb / Q)
Streamflow Coefficient	CE	Indicator of the amount of precipitation that became streamflow in a basin (CE = Q / P)

References

• Bamler, R. (1999). The SRTM mission: A world-wide 30m resolution DEM from SAR interferometry in 11 days. Photogrammetric Week.

• Bontemps, S., Defourny, P., Radoux, J., Van Bogaert, E., Lamarche, C., Achard, F., et al. (2013). Consistent Global Land Cover Maps for Climate Modeling Communities: Current Achievements of the ESA’s Land Cover CCI. In ESA Living Planet Symposium.

• Funk, C., Peterson, P., Landsfeld, M., Pedreros, D., Verdin, J., Shukla, S., et al. (2015). The climate hazards infrared precipitation with stations - A new environmental record for monitoring extremes. Scientific Data, 2, 1–21. https://doi.org/10.1038/sdata.2015.66

• Hengl, T., De Jesus, J. M., Heuvelink, G. B. M., Gonzalez, M. R., Kilibarda, M., Blagotić, A., et al. (2017). SoilGrids250m: Global gridded soil information based on machine learning. PLoS ONE. https://doi.org/10.1371/journal.pone.0169748

• Lindström, G., Pers, C., Rosberg, J., Strömqvist, J., & Berit, A. (2010). Development and testing of the HYPE (Hydrological Predictions for the Environment) water quality model for different spatial scales. Hydrology Research, 4(41.3), 295–319. https://doi.org/10.2166/nh.2010.007

• Mu, Q., Zhao, M. and Running, S. W. (2013). MODIS Global Terrestrial Evapotranspiration (ET) Product (MOD16A2/A3), Algorithm Theor. Basis Doc.