This archive contains the CSV files supporting the response to a community comment:

“Is it possible to provide the list of selected 218 catchments from the CAMELS_AUS dataset and the median values of performance metrics for all the models?”

The files provide Nash–Sutcliffe Efficiency (NSE) results used in the paper:
Better continental-scale streamflow predictions for Australia: LSTM as a land surface model post-processor and standalone hydrological model
(Shokri et al., 2025, in review).


File Structure:

Each CSV corresponds to a model and cross-validation setup described in Section 2.5 of the paper.
The filenames indicate the model type and experiment (TooS = Temporal OOS, SooS = Spatial OOS, TSooS = Spatiotemporal OOS).

File			Description
LSTM_qc_toos_NSE.csv	LSTM-QC under Temporal Out-of-Sample test
LSTM_qc_soos_NSE.csv	LSTM-QC under Spatial Out-of-Sample test
LSTM_qc_tsoos_NSE.csv	LSTM-QC under Spatiotemporal Out-of-Sample test
LSTM_c_toos_NSE.csv	LSTM-C under Temporal Out-of-Sample test
LSTM_c_soos_NSE.csv	LSTM-C under Spatial Out-of-Sample test
LSTM_c_tsoos_NSE.csv	LSTM-C under Spatiotemporal Out-of-Sample test
gr4j_toos_NSE.csv	GR4J under Temporal Out-of-Sample test
gr4j_soos_NSE.csv	GR4J under Spatial Out-of-Sample test
gr4j_tsoos_NSE.csv	GR4J under Spatiotemporal Out-of-Sample test
awra_NSE.csv		AWRA-L baseline NSE values across 218 catchments


Catchment Information:

All 218 catchments are from the CAMELS-AUS dataset (Fowler et al., 2021).


Each CSV includes:

station_id – CAMELS-AUS catchment ID
NSE – Nash–Sutcliffe Efficiency