This month we interview Kazuhito Ichii — a Professor in the Center for Environmental Remote Sensing (CEReS), Chiba University, Japan. Kazu recently represented AsiaFlux to attend the FLUXNET 2017 workshop, and gave a presentation on the updates of AsiaFlux Network.
Please tell us briefly about yourself and your lab.
I recently joined Center for Environmental Remote Sensing (CEReS), Chiba University, Japan as a professor, and launched a new group in this April. I received a PhD degree of Earth Science from Nagoya University, Japan in 2002, and worked as a post-doc in NASA Ames Research Center, USA under the supervision of Dr. Ramakrishna Nemani, an associate professor in Fukushima University, Japan, and a senior scientist in Japan Agency for Marine-Earth Science and Technology (JAMSTEC).
My background of research is earth system modeling and remote sensing, slightly different from FLUXNET communities. My previous work more focused on global scale earth system modeling and continental scale monitoring using remote sensing (without ground observation based validation at that time). As my PhD works, I developed my own simple zero-dimensional earth system model including atmosphere, ocean, and land, and conducted vegetation change analysis using time-series NOAA-AVHRR at global scale as PhD works.
Currently, my works (including my group) mainly focus on understandings and quantification of land-atmosphere greenhouse gas exchanges from country to global scales. Our approach is “integration” of available information such as FLUXNET, AsiaFlux, remote sensing, ecosystem models, and atmospheric inversions. I am working with many collaborators/friends in AsiaFlux, FLUXNET, FLUXCOM, remote sensing, terrestrial modeling, and earth system modeling to try to expand and fulfill our ideas.
What was your path towards measuring/modeling fluxes between land and the atmosphere?
Many friends and seniors brought me into the flux world!! As I mentioned above, my PhD work is not directly related to land-atmosphere fluxes. The first turning point was in 2002, just after I received PhD. Two of my seniors (Prof. Kenlo N. Nasahara, Tsukuba Univ. and Prof. Atsushi Higuchi, Chiba Univ.) connected me to Dr. Ramakrishna Nemani to stay in University of Montana during summer. At that time, I had interests in terrestrial ecosystem modeling to interpret our results of vegetation changes detected by NOAA AVHRR. During the stay (two months in summer), I learned a lot on Biome-BGC model and also found many problems. So, I decided to use eddy-covariance observation data to test the models. The second turning point is after I came back to Japan and meet JapanFlux and AsiaFlux members. Several colleagues (Drs. Nobuko Saigusa and Ryuichi Hirata, National Institute for Environmental Studies, Japan) in Japan introduced me to work with JapanFlux and AsiaFlux members. In particular, I served as a model-upscaling group leader of CarboEastAsia program (www.carboeastasia.org), international joint project among China, Korea, and Japan, and met many researchers in the three countries. This work really fit to me, and I could expand my experience to international collaboration. Prof. Joon Kim, Seoul National University, is a great mentor to conduct the model comparison project, and we became a great collaborator.
What are some of the challenges you are facing modeling the fluxes and making defensible prediction?
Before I used terrestrial ecosystem models, I expected that the models were perfect. However, when I used these models, I found models are immature. The first challenge was to collect enough observation data to test models. We had insufficient observation network and observation parameters in about 15 years ago. Therefore observation data were not sufficient to constrain the model. In recent years, data-driven models were greatly developed and these models provides a new dataset of terrestrial fluxes. Therefore, we can refine models using newly available observations now. Second challenges are way of model-data integration. Current terrestrial models have too much complexity and too many parameters, therefore, model-data fusion is computationally expensive. Although new techniques of model-data fusion were developed, these studies are mostly at site-level studies. Large-scale application of model-data fusion is very challenging. Third challenge is there’re too many different models, and we don’t know which one is the best (or better) for each specific applications. I believe it’s a time to re-organize existing models to develop more synthesized, harmonized and well-evaluated models.
What is your unique niche with regards to Asiaflux and Fluxnet? In other words, what additional or distinct ideas, skills or measurements to you bring to the community to add to our intellectual diversity and breadth of data?
I am working on integration of eddy-covariance data, remote sensing data, modeling, and inverse estimations, and I believe this is my unique niche. In fact, I don’t have any specific technique of measurement and I am not a model developer. Therefore, integration (bringing many knowledges to find new things) is the only my path to survive in this research field. Fortunately, I like computer programming and data analysis. Therefore, I decided myself to focus on integration.
What are some of the most exciting findings from your work so far?
Two works so far. One is to demonstrate how vegetation related parameters which are difficult for direct observation using observation data (e.g. Ichii et al. 2007 and 2009). In other words, inverse estimation of unknown parameters using model-data fusion techniques. In particular, Ichii et al. (2009) AgForMet paper estimated vegetation rooting depth in California using empirically upscaled ET map. We inversely estimate rooting depth using Biome-BGC model and upscaled ET seasonal variation based on the ET variations during dry season. Other exciting work is to link FLUXNET data to earth system model and future projection (Suzuki and Ichii, 2010; Tellus B). We used about 50 sites eddy-covariance data, and tested and improved terrestrial submodel in ESM, and tested the impact on the future climate projection. It was a rough, but very innovative study.
As an senior scientist and professor, what advice do you have for younger scientists and students?
Knock the doors of other groups and don’t hesitate to discuss with other scientists. I like to visit and work together with other groups. In the past, I visited several groups for discussion and collaboration, and all of them were successful.
Fundamentally, having basic skills and backgrounds are very important. In this study field, I believe data-analysis technique (e.g. computer programming), basic mathematics and statistics, physics, and English are minimum requirements. If you have them (plus good publications), you can work everywhere, at least as a postdoc researcher. Many PIs really needs such a person.
Running a lab, what is your style of leadership and mentoring?; and how have past mentors influenced this?
I am always trying to work very closely with students and postdocs to run the project. In the first stage of graduate course, I believe students need to discuss a lot with supervisor even for the basic questions and their future careers.
I learned too many things from Dr. Rama Nemani, NASA Ames Research Center. At first I was too much shocked, since I found many research experiences during PhD were too immature. As one experience, in my postdoc work, I finished in a week and shared one manuscript with Rama, however, he told me that paper cannot be finished within a week and I should improve it more. In the preparation process of papers in my postdoc period, I realized that preparation of manuscript is very difficult and time-consuming tasks. Rama always has comprehensive, impressive and simple ideas – I believe simplicity is very important -, and also led me with very simple instructions. Estimation of rooting depth from remote sensing data and ecosystem models are also emerged through the discussion with Rama. He just came to us and explain his idea briefly, and we deeply think about how to achieve, and finally work had done.
Finally, our group needs good, self-motivated graduate students. Please contact me if you have interests !
More to know about Kazuhito and his group :
Group URL: http://ichiilab.weebly.com/
Ichii K. et al. (2017) New data-driven estimation of terrestrial CO2 fluxes in Asia using a standardized database of eddy covariance measurements, remote sensing data, and support vector regression. Journal of Geophysical Research Biogeosciences. 122, 767-795, doi:10.1002/2016JG003640.
Suzuki T., and Ichii K. (2010) Evaluation of a terrestrial carbon cycle submodel in an earth system model using networks of eddy covariance observations, Tellus 62B, 729–742.
Ichii K. et al. (2009) Refinement of rooting depths using satellite-based evapotranspiration seasonality for ecosystem modeling in California. Agricultural and Forest Meteorology, 149, 1907-1918.
Ichii K. et al. (2007) Constraining rooting depths in tropical rainforests using satellite data and ecosystem modeling for accurate simulation of GPP seasonality, Global Change Biology, 13, 67-77.