First 6 months - Year One
The project commenced on 1st July 2004. The first project planning workshop was held in Bangkok, Thailand from 18th to 24th July 2004. The project leader, Dr Yahya Abawi and the whole Indonesian team participated at the meeting. Restrictions on travel have delayed some key meetings between the Australian and Indonesian teams. However, communication via email, internet, and telephone has enabled progress to be made.
During the reporting period, the project team in Lombok have met regularly and have conducted several workshops on seasonal climate forecasts and FLOWCAST software. A program to train the government officers on understanding FLOWCAST and strategic farming decisions is expected to be conducted between March and June 2005. The training material has already been prepared. To disseminate the project results the expansion of existing linkages to prospective final users such as government institutions, commercial companies and farmers is planned.
The development of the Linear Programming Model (LP) for optimising crop decisions is progressing. The data for refining the LP model and further calibration of the IQQM (Integrated Quantity Quality Model) are being collected. Field visits to Sengkol and Mujur have been conducted to establish 500-1000 ha pilot field trials in irrigated and rain-fed lowland areas in southern Lombok in 2005. Interviews with local farmers have shown there is potential to grow long bean, cucumbers and chillies as the second (rice is the first) and third crops in the region.
To better communicate the progress of the project among the Indonesian team and stakeholders, a project Newsletter or Bulletin has been created. It is published electronically and in hardcopy (limited copies) in both English and Bahasa Indonesia 6 times a year.
Hardcopies of the bulletin will be circulated to universities and Direktorat Balitbang Pertanian Jakarta and government institutions in West Nusa Tenggara.
During the reporting period, considerable effort has been made by the Australian team to develop new and improved tools to incorporate into the CropOptimiser software package. This package provides a graphical user interface to the LP model to optimise different cropping strategies for a range of seasons and climate types. New tools that have been developed include a GIS based user-interface allowing direct interaction with the model through the map itself. An XML-based data mechanism has been included to incorporate a range of user-definable crop templates which can simulate a range of crops such as rice, chilli, vegetables and their combinations. The chart outputting mechanism of the original version of the package has been completely rewritten by using up-to-date software engineering technology, resulting in a much simpler, yet more powerful user-interface to analyse the simulation results and produce an output report. The release of a Beta version of CropOptimiser is expected towards the end of June 2005. Considerable revision has also been made to the IQQM Lombok model. The model configuration of IQQM GUI (Graphical User Interface) version has been completed, which gives the functionality to configure the river systems on GIS-based geographic display. The calibrations have been extended to other sub-systems such as Nyurbayah, Mencongah, Majeli, and Repok Pancor in Jangkok rivers system, the upstream of Mataram (i.e. at Bugbug), and Hilir, Hulu, Batu, Katon, Mujur in Jurang-Satte rivers system. Sacramento models were developed for estimating the flows in Jangkok and Sesaot river. Currently, the meteorological data, hydrological data, and irrigation diversion data that are needed for configuring and calibrating other sections in the system are being gathered in Lombok. The travel by Dr Xike Zhang to Lombok to obtain the data sets for further configuration and calibration of IQQM model is planned for late April 2005.

For 2005, the overall progress of the project can be summarised under the following sub-headings:

1. Development of water allocation model (IQQM)
IQQM (Integrated Quantity and Quality Modelling) is a computer simulation model developed by the Department of Infrastructure, Planning and Natural Resources, NSW which can simulate river flow, flow routings, water allocation and water use at a catchment scale. In this project, the model is configured to simulate the river-irrigation system in Lombok. Considerable effort has been made by the Indonesian team to collect and digitise a variety of data including meteorological, streamflow and cropping data in the past year. The data have been quality-checked and patched where necessary using statistical methods by the Australian team. Model verification and configuration has progressed well through emails and face-to-face discussions with the Indonesian partners. At the time of this report the IQQM calibration including streamflow and diversion calibration has been completed for the whole system except for two irrigation nodes where data is not available yet. The data collection and fine-tuning of the model is continuing.

2. Generation of long-term meteorological and hydrological data
The aim of the project is to use climate forecasts to improve the management of water resources and irrigation systems in Lombok. To achieve this we need to run model simulations over a long historical data set to capture variability in climate and other management decisions. Long-term (more than 50 years) of meteorological and riverflow data are needed to simulate different scenarios in the IQQM model. Data on rainfall, temperature, solar radiation and riverflow are either limited in length or discontinuous for most of the stations which represent the different catchments in Lombok. A number of models such as Sacramento, IHACRES (identification of unit hydrographs and component flows from rainfall, evaporation and streaflow data), Simhyd and AWBM models were assessed and used to simulate (generate) river inflow and other meteorological data for about 50 years. So far dis-aggregation of daily rainfall (from recorded monthly data) and the generation of climate data for 1950-2000+ have been completed for 12 stations which cover the west and central part of Lombok. The data generation for eastern Lombok is continuing.

3. Development of Flowcast software
Flowcast is developed as a tool to assess the effect of ENSO and other climate predictors on rainfall, streamflow and crop production. Flowcast is designed to provide the necessary interface between simulation output from the IQQM model and climate variables such as the Sea Surface Temperature (SST) and the Southern Oscillation Index (SOI). Software engineering has seen the continuing development of FlowCast and the new version of FlowCast combines many new features of graphical output analysis along with improved forecast and skill testing capabilities. A prototype version of FlowCast has now been completed, and is currently undergoing testing and debugging.

4. Development of Linear Programming model and CropOptimiser
Linear Programming (LP) and CropOptimiser DSS are designed to use linear optimisation to optimise choice of crop for given climate, land, water, and system constraints. During 2005, work on LP model has focused on data collection, calibration and validation. This work forms the basis of a PhD study by Ir. Ismail Yasin of the University of Mataram.

Considerable progress has been made by the Australian team in the development of the CropOptimiser software. CropOptimiser is developed to combine LP programming and model output through graphical user interface. This will provide greater flexibility and power to generate and display outputs. A prototype will be available by May 2006.

5. Information dissemination and capacity building
During this reporting period, the project personnel from Australia made two trips to Lombok. Activities included field visits, data acquisition, training on IQQM modelling, PhD supervision and presentations in workshops. Field visits and farmer interviews have been conducted by the Indonesian collaborators to investigate and evaluate framers' acceptance of the information and knowledge developed in the project. A report is in preparation. A farmer group has been established in the Batujai irrigation area and field days have been conducted to promote the use of water harvesting and alternative cropping systems in this drought-prone area. Attendees at these workshops included local farmers, village leaders and village chiefs.

6. Water Balance Study in PRB cropping systems of southern Lombok
Work on the water balance study of permanent raised beds started in the latter half of the year. A detailed analysis of water balance in southern Lombok has been conducted by the Australian team using a water balance model (HowLeaky). Preliminary water balance simulations for Mankung show that approximately 130 mm per annum of excess water may be harvested from permanent raised beds (PRB) that could be used to supplement irrigation water during the second season. However, inter-annual variability is high and further analysis will be carried out using long term climate data to determine the impact of extended drought and ENSO on water harvesting and implications for infrastructure to store and re-use the excess water.

The overall progress of the project can be summarised under the following sub-headings:

1. Configuration and Validation of the IQQM model (Integrated Quantity and Quality Model) for Lombok.
During 2006, the configuration and validation of the IQQM simulation model was completed. IQQM is used to simulate catchment-scale river flows, flow-routings, water allocations and water-use. These outputs can then be used for input into the decision support systems (FlowCast, CropOptimiser) to forecast irrigation water allocations and improve decision-making on water diversion and crop-management (e.g. planting area, crop type) in Lombok. To capture the long term variability in climate, setting up IQQM requires long-term meteorological and river-inflow data. This involved synthesising daily rainfall and temperature data from long-term monthly records using the WeatherMan package. These data were then used to calibrate the IHACRES rainfall-runoff model against short-term historical streamflow records to generate long-term daily river inflows and sub-catchment residual inflows. Using these generated data as inputs, IQQM was calibrated against short-term (1995-2000) available streamflow, water diversion and crop requirement data. The model can now generate more than 50 years of daily streamflow data for the 57 irrigation areas in the Lombok irrigation area. This work was presented at ANZ Climate Forum in Canberra, Australia in 2006. A study of streamflow forecasting in the Jankok River has also been completed and presented at the 2nd International Rice Congress in New Delhi, India in 2006. Simulating and forecasting of streamflow in other rivers is continuing.

2. Development of Decision support systems
Development of the key decision support software systems has continued, along with a new program - Pump Test Analyser:
CropOptimiser is a decision support tool for optimising cropping patterns for different climate, land, water and system constraints. The first full prototype version of the software that includes the Linear Programming (LP) model was developed during the year. The LP model was converted into C++ from its EXCEL prototype form and embedded into the highly graphical interface, and linked to a custom-designed editor for inputting climatic, hydrologic, agronomic and social constraints.
FlowCast is decision support software for assessing the effects of climate on rainfall, streamflow and crop production. It provides the necessary connection between the IQQM model outputs and climate variables such as ENSO to develop seasonal climate outlooks, and forecasts of water availability. During 2006, a new graphical user interface was developed to minimise program complexity combining both spatial and point-based outputs. New analyses including LEPS (Linear Error in Probability Space), real-time forecast verification, Receiver Operating Characteristics (ROC) testing, interactive probability and analogue outputs have been added.
HowLeaky is a landuse evaluation tool based upon a simple one-dimensional water balance model for evaluating different soils, vegetation, tillage, irrigation, pesticide and nutrient scenarios. During 2006, it was modified and updated to include specific crops, soils and management scenarios (including raised-bed and traditional agricultural practices) for Lombok.
Pump Test Analyser was developed in 2006 to analyse borehole pumping tests undertaken in Lombok during 2005/2006. The program inputs drawdown and recovery data and calculates maximum pumping rates and specific capacity for a user-specified time-scale in the form of a customizable report. This software will be used in a study to determine the potential groundwater resources for irrigation in Southern Lombok.

3. Information dissemination and capacity building
During this reporting period, project personnel from Australia made two trips to Lombok including the ACIAR project review in September. Two training courses were held including training of the staff from BMG in the use of seasonal climate forecasting (SCF), and a 5-day IQQM training course for the Public Works and UNRAM teams. The Indonesian team also conducted a workshop on 19-20 October 2006 on the importance of the project on irrigation and agricultural strategies at district and subdistrict levels in collaboration with the Dinas Pertanian (Department of Agriculture) program.

Dissemination of project outputs was introduced to farmer groups, field officers and government agencies in Batujai and Kawo. However, a significant opportunity was lost to communicate to farmers about the consequences of the 2006 El Nino, which resulted in a delayed onset to the monsoon season and caused sowing failures in dryland areas. This can be used as a learning experience for the future communication strategy which includes strengthening the climate knowledge and information exchange between the governing and extension bodies.

4. Shallow ground-water study
A survey was carried out to understand the characteristics of shallow ground water in 197 shallow wells in Southern Lombok. Measurements included depth to groundwater, draw-down and recovery of wells, and water quality (NO3, EC, TDS, and pH). Results showed that the depth of the water table is highly correlated to the ground surface-elevation, transmisivity varies seasonally, while the quality of ground water is good and safe for domestic purposes and agriculture. Further work is continuing to assess the extent of groundwater for supplementary irrigation.

5. Water balance study in PRB cropping systems of southern Lombok
A water balance study of permanent raised beds (PRB) was completed at the start of the year using the HowLeaky software for local crops and soils. Results show that approximately 130mm of excess water may be available with PRB from Phase 1 (Nov-Mar) although inter-annual variability is extreme. Although carryover is insufficient to fully irrigate a similar area in Phase 2 (Mar-June), restricting irrigation to early crop development (80mm) may provide satisfactory yields providing enough storage is available. Further work is continuing to assess the feasibility and economics of such storage systems. Mulching the PRB had little effect on the crop irrigation requirement (10-15mm) because while it reduces evaporation, it increases deep drainage with minimal net benefit.

Background
This project aims to use seasonal climate forecasting to improve water resource and irrigation systems management in Lombok, Indonesia. Agricultural production in Lombok is predominantly rice-based and is largely influenced by high climate variability that, in part, is associated with the El Nino Southern Oscillation (ENSO) phenomenon. Therefore, using ENSO-based seasonal climate forecasts (SCFs) ahead of the growing season can potentially improve agricultural productivity in favourable seasons and reduce the risk of crop losses in dry years through tactical adjustment of crops and water allocation. Key objectives of the project are to:
Collect, synthesise, model and collate hydrologic and climatic data for integration into a climate-based decision support system;
Develop decision support tools for optimising choice of crop, crop-area and irrigation water allocation based on seasonal climate information;
Promote SCF-based planning amongst irrigators, government officials and community leaders; and
Build local capacity in the development and operational use of decision support systems.
Summary of progress, June 2007 to May 2008
The scientific components of the project are now mostly complete. However, collation and dissemination of results and outputs remain ongoing, with few tangible impacts observable at the community level. This reflects the complexity and the "systems" nature of the project, which relies on the integration of many completed components to produce useful information for dissemination to stakeholders. Progress in this reporting period has mainly focused on finalisation of the scientific components, and dissemination of the project outputs to the government and provincial agencies with several workshops and training activities conducted both in Australia and Indonesia.
Decision support development
Significant work was undertaken in the previous reporting periods in data collection (meteorological, hydrologic and agronomic), patching and synthesis for input and calibration of the hydrologic model (IQQM). Modelling outputs include simulated catchment-scale river flows (50-year daily streamflow), flow-diversions, and water allocations and usage at fifty-seven irrigation weirs in southern Lombok. The hydrological model has now been configured to simulate "historical" water allocation for two demand scenarios (existing cropping pattern and maximum planted area). Three stochastic realisations of streamflow were used in each analysis to capture model uncertainty.
These data are now available for input into the FlowCast and CropOptimiser decision support software for tactical decision-making. FlowCast was developed to generate probabilistic forecasts of rainfall, streamflow and water allocation based on key climate predictors. CropOptimiser (a linear programming (LP) model) uses probabilistic forecast distributions from FlowCast to optimise choice of crop and planted-area for different seasons and climate types. Prototype versions of these software packages were developed in previous reporting periods. However, continued development over this period has now seen operational versions of both software packages released.
FlowCast was significantly modified during the reporting period to simplify its operation, and to enhance its station and spatial analysis capabilities. Powerful forecast analysis tools were added, and the graphical user interface was redesigned to minimise the operational requirements of the software, and to hide the complexity inherent in a state-of-the-art technology. FlowCast was formally released to the Indonesian collaborators at a national meeting of the Indonesian National Agency for Meteorology and Geophysics (BMG) in Jakarta in April 2008.
CropOptimiser was also extensively modified to produce the first fully functional version. The algorithms used in FlowCast to calculate forecast distributions have now been directly incorporated into CropOptimiser to minimise the operations required to generate output. This now facilitates direct integration of IQQM outputs into CropOptimiser allowing for testing of different scenarios based on alternative IQQM simulations. Data management algorithms have been developed to aggregate IQQM node outputs based on defined irrigation sub-area configurations, and to average different climate realisations. New reporting capabilities have also been added and the software debugged and tested.
Another accomplishment of the project was the creation of an internet-based database system to manage and store the vast amounts of collected input data. This is potentially one of the most important outputs of this project that can be used for current and future research dealing with climate and water resources management in Lombok. It was envisaged that a similar database would also be generated containing simulation results covering a range of climate scenarios, allocation decisions and planting options. This task has been delayed due to organisational changes in Australia and difficulties in recruitment of staff which ensued. Nevertheless, modifications made to the decision support tools during 2007 will greatly facilitate this task. We will be seeking approval to undertake this work after the final reporting period using unspent project funds.
A further objective of this project negotiated with ACIAR in July 2006, in collaboration with the La Trobe University (SMCN/1999/005), was to conduct a water balance modelling study to assess the potential of water harvesting from the raised-bed systems in southern Lombok, using the HowLeaky software package. Earlier work involved modifying the HowLeaky simulation code and creating parameter-files for conditions in Lombok. In this reporting period, HowLeaky was rewritten to simplify its operation and to directly simulate multiple cropping cycles, and run batch simulations for a detailed sensitivity analysis of different parameters. Outputs from HowLeaky were used in FlowCast to assess the impact of seasonal climatic variability on field-runoff in the southern Lombok region. This study is ongoing and will be completed by the final project review, with an additional site studied.
Another requirement of the project is to assess the feasibility of using groundwater as a supplementary source of irrigation in the drier areas of southern Lombok. A survey of 197 wells and multiple pumping tests have been carried out, to determine safe groundwater yield and water quality in the region. This data was reanalysed in this reporting period to better quantify the results. Results showed that shallow groundwater resources are suitable for agriculture but are limited in capacity and can only be used as a supplementary source using "hand-watering" methods.
Consultation, information dissemination and capacity building
Information exchanges continued throughout this period in both Australia and Indonesia. There now appears to be a good understanding of project functionality and responsibilities between key partners, with relative success in scientific capacity building. Unfortunately, at this stage in the project, information dissemination to key stakeholders is still limited, focusing on promoting the theory and advantages of the technology, rather than introducing effective changes into the local communities. It is unlikely that any practical outcomes of this research were implemented at the farm level to influence decision-making, despite attempts to do so.
These issues were addressed in August 2007 through a workshop involving the Australian team and key Indonesian scientists in Toowoomba (Australia). A communication strategy and detailed plan of local capacity building were developed for disseminating project outputs to farmer groups, field officers and government agencies.
In Indonesia, efforts in dissemination of information focused on three different levels including: scientific academics and government officials; field extension and water gate managers; and village leaders and farmers. National workshops were held in Jakarta (PERHIMPI) and Bali (related to United Nations Framework Convention on Climate Change (UNFCCC)) where papers were presented on the impacts of climate variability on agriculture and the use of FlowCast as a tool for seasonal climate forecasting in Lombok. At the regional level, workshops and training programs were conducted to increase officer understanding on climate variability to introduce the concepts of decision support systems. Informal meetings and consultations took place with the head of Bappeda Provincial Office and the head of Department of Agriculture at District of Central Lombok, on adopting the decision support tools for strategic cropping management. Officers from Dinas Pertanian (Department of Agriculture) conducted field days and workshops for local farmers on climate, while the Indonesian project team also visited some farmers' group associations to lecture on the local impacts of climate variability and climate change.