This document discusses methods for assessing water resources in areas with little available data. It begins by examining approaches to estimating annual runoff in ungauged basins using empirical statistical relationships between runoff and factors like rainfall, elevation, and temperature. These relationships have been developed and tested for regions in Italy. The document then explores using climatic indices and variables like net radiation, evapotranspiration, and the Budyko index to characterize climates and their relationship to runoff with limited data. Methods are presented for reconstructing temperature patterns from elevation, latitude and other variables. The potential of using satellite-derived NDVI data as a proxy for rainfall is also discussed. The document describes an ongoing project in Italy to compile hydro
CLAPS - WATER RESOURCES ASSESSMENT IN DATA-SCARCE AREAS
1. Water resources assessment in
data-scarce areas
Pierluigi Claps - Politecnico di Torino and HydroAid
pierluigi.claps@polito.it
www.idrologia.polito.it/~claps
4. What does UB mean?
(Really totally Ungauged?)
NO Runoff
but
what about climatic data?
5. ANNUAL RUNOFF ESTIMATION IN UNGAUGED BASINS (1)
Case study: Basilicata Region
(10000 km2 - 22 gauged test basins)
6. Empirical Statistical Estimation of Dm
(Basilicata)
1
Dm = -24,8 + 4.37 ⋅ ln Pm + 0,0028 ⋅ z
3
OK, R2=0.9552 but, how to adopt the relation outside
the calibration region?
Dm = average annual runoff (mm)
Pm = average annual rainfall (mm)
z = average basin elevation (m a.s.l.)
7. ANNUAL RUNOFF ESTIMATION IN UNGAUGED BASINS (2)
Case study: Piemonte Region
(25000 km2 - 47 gauged test basins)
Very heterogeneous region in climate and
morphology
13. a simple question:
• what to do when calibration basins are very few?
• is average climate related to annual runoff?
• are there other useful physical information?
8
14. IS AVERAGE CLIMATE RELATED TO ANNUAL RUNOFF ?
CLIMATIC INDICES can be computed using
Rainfall and ...
… Temperature (Emberger, 1955):
…Potential evapotranspiration (Thornthwaite, 1948)
…Solar radiation (Budyko, 1956)
Rn
I=
λ⋅P
ETp = average annual potential evapotranspiration (mm)
P = average annual rainfall (mm)
M = mean temperature of the hottest month (K) Rn = average annual net radiation (MJ/m2)
m = mean temperature of the coldest month (K) λ = latent heat of vaporization (MJ kg-1)
€
15. Selection of the minimum necessary
information, for use in data-scarce areas
• are Climatic Indices meaningful?
• are they related to runoff?
• how much information is really necessary to compute them?
16. climatic variables in Basilicata (partially reconstructed)
NE
MEAN ANNUAL RAINFALL MEAN ANNUAL TEMPERATURE
41 2400
2200
2000
40.5 1800
Latitudine
1600
1400
40 1200
1000
800
39.5 600
14.5 15 15.5 16 16.5 17
Longitudine
MEAN ANNUAL PET NET RADIATION
17. 1. Comparison of different climatic indices (Claps & Mancino, 2002)
HUMID HUMID
ARID ARID
ARID
(x-µ)/σ HUMID
19. 3. Evaluation of the minimum necessary amount of information
BASIC VARIABLES CLIMATIC VARIABLES
Terrain Elevation
Temperature
Latitude Net Radiation
Average Cloudiness factor
(relative eliophany)
Precipitation
Evapotraspiration
20. 3. Evaluation of the minimum necessary amount of information
BASIC VARIABLES CLIMATIC VARIABLES
Terrain Elevation
Temperature
Latitude Net Radiation
Average Cloudiness factor
(relative eliophany)
Precipitation
Evapotraspiration
21. 3. Evaluation of the minimum necessary amount of information
BASIC VARIABLES CLIMATIC VARIABLES
Terrain Elevation
Temperature
Latitude Net Radiation
Average Cloudiness factor
(relative eliophany)
Precipitation
Evapotraspiration
22. 3. Evaluation of the minimum necessary amount of information
BASIC VARIABLES CLIMATIC VARIABLES
Terrain Elevation
Temperature
Latitude Net Radiation
Average Cloudiness factor
(relative eliophany)
Precipitation
Evapotraspiration
23. 3. Evaluation of the minimum necessary amount of information
BASIC VARIABLES CLIMATIC VARIABLES
Terrain Elevation
Temperature
Latitude Net Radiation
Average Cloudiness factor
(relative eliophany)
Precipitation
Evapotraspiration
24. Empirical T(z,Lat) estimation
(Claps and Sileo, 2001)
stations in Southern Italy
Mean annual Temperature (°C)
Mean monthly Temperature
25. Empirical T(z,Lat) estimation
(Claps and Sileo, 2001) 25
stations in Southern Italy
Temperatura media mensile °C
20
15
Mean annual Temperature (°C) 10
5
0
Gen Feb Mar Apr Mag Giu Lug Ago Set Ott Nov Dic
Pescopagano
25
Mean monthly Temperature
Temperatura media mensile °C
20
15
10
5
0
Gen Feb Mar Apr Mag Giu Lug Ago Set Ott Nov Dic
Melfi
30
Temperatura media mensile °C
25
20
15
10
5
0
Gen Feb Mar Apr Mag Giu Lug Ago Set Ott Nov Dic
Policoro
26. Empirical T(z,Lat) estimation
(Claps and Sileo, 2001) 25
stations in Southern Italy
Temperatura media mensile °C
20
15
Mean annual Temperature (°C) 10
5
0
Gen Feb Mar Apr Mag Giu Lug Ago Set Ott Nov Dic
Pescopagano
25
Mean monthly Temperature
Temperatura media mensile °C
20
15
10
5
0
Gen Feb Mar Apr Mag Giu Lug Ago Set Ott Nov Dic
Melfi
30
Temperatura media mensile °C
25
20
15
10
5
Relations affected by the scale 0
Gen Feb Mar Apr Mag Giu Lug Ago Set Ott Nov Dic
Policoro
of the analysis?
35. From Satellite images (GIMMS - http://glcf.umiacs.umd.edu)
Normalized Difference Vegetation Index (NDVI)
Depends on measures of reflectance in the Visible (RVIS) and in the Near-Infrared (RNIR) :
Claps and Laguardia, 2004
43. ~ 500 basins with runoff data
Maximum annual instantaneous and
daily discharge; several daily runoff
time series, etc.
~ 6000 rainfall stations
Maximum annual daily rainfall, max
annual rainfall in 1-24 hrs (40% of the
stations), etc.
44. The Information System of the Italian basins
- fully open source (grass-postgres-openI)
- GIS raster and vector database-compliant
- compatible with the CUAHSI information system
48. seasonality of NDVI vs seasonality of runoff
• Average NDVI (16 days-values) for each catchment
• Distance between the NDVI in 2 catchments d = mean|NDVI1,i – NDVI2,i|
• Distances for each pairs of curves distance matrix DNDVI
• Analugous distance for monthly streamflow regime curve
All NDVI regimes (431
MOPEX catchments)
|NDVI1,i–
49. NDVI in Italy (awaiting for application)
A parameters of the first harmonic
F (months)
50. thanks, and come to
Turin!
papers on the topic available at:
37
http://www.idrologia.polito.it/risorseidriche/download.html
Hinweis der Redaktion
Calcolo delle matrice delle distanze per:
Curve dei regimi idrometrici mensili
Curve dei regimi relativi al NDVI (media sull’area del bacino dei pixel delle immagini modis) con frequenza 16 giorni
di 431 bacini del USGS inclusi nel database MOPEX (Model Parameter Estimation Experiment) (ftp://hydrology.nws.noaa.gov/pub/gcip/mopex/US_Data/)